autotouch-cli 0.2.85__tar.gz → 0.2.87__tar.gz

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: autotouch-cli
-Version: 0.2.85
+Version: 0.2.87
 Summary: Autotouch Smart Table CLI
 Project-URL: Homepage, https://app.autotouch.ai
 Project-URL: Documentation, https://github.com/nicolonic/autotouch_main/tree/main/docs/research-table/reference
@@ -132,7 +132,7 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 - Inspect rows: `autotouch rows list`, `autotouch rows get`
 - Inspect one cell: `autotouch cells get`
 - Create a workflow column: `autotouch columns recipe`, `autotouch columns create`
-- Run neural search for niche/specific company or people discovery: `autotouch search companies`, `autotouch search people`
+- Run Exa Company Search for basic company discovery, competitors, and lookalikes: `autotouch search companies`, `autotouch search similar-companies`
 - Inspect list-build inputs: `autotouch list-build inputs`
 - Build durable company and lead lists on Smart Table workers: `autotouch list-build companies`, `autotouch list-build leads`
 - Check connected LinkedIn account status/limits: `autotouch linkedin status`, `autotouch linkedin limits`
@@ -146,13 +146,17 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 
 Autotouch has two different company/people discovery paths.
 
-Use `autotouch search companies` and `autotouch search people` for neural search when the target is niche, semantic, or hard to express with structured filters. Examples: "AI workflow platforms for healthcare", "AI workflow startups selling to law firms", or "RevOps operators at PLG SaaS companies". Neural company/people search returns at most 10 results per API call and costs 1 credit per API call.
+Use `autotouch search companies` for Exa Company Search when the target is basic company discovery, competitors, lookalikes, or a semantic company search. Put company constraints directly in natural language: industry, geography, funding, headcount, technology, founded date, and similarity. Examples: "AI workflow platforms for healthcare", "US education software startups focused on AI math tutoring with 1-50 employees", or "companies like Rippling for HR and workforce management". Exa company search returns up to 100 results per API call and costs 1 credit per 10 returned results.
 
-Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
+Use `autotouch search similar-companies` as a convenience wrapper for Exa Company Search similarity. Prefer `--query` for Exa's documented semantic patterns, such as "companies like Notion", "competitors of Stripe", or "companies working on making space travel cheaper". The older seed fields `company_name`, `company_domain`, and `company_description` remain available when you want the CLI/API to compose that query; it does not use Exa `/findSimilar`.
+
+Use `autotouch search people` for provider-hidden people search when the target is niche, semantic, or hard to express with structured filters. People search currently returns at most 10 results per API call and costs 1 credit per API call.
+
+Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable LinkedIn/Sales Navigator list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
 
 For structured company builds, start with industry IDs plus geo/company-size filters. User-supplied `keywords` are optional when structured filters are present; use them only to refine, disambiguate, or recover hard-to-classify targets instead of carrying the whole target definition.
 
-For account-first prospecting, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use neural search as a default pre-step for every list build; use it when the user's target is genuinely semantic or niche.
+For account-first prospecting from LinkedIn/Sales Navigator, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use Exa company search as a default pre-step for every LinkedIn list build; use it when the user's target is genuinely semantic, niche, competitor-based, or not meant to be LinkedIn-sourced.
 
 ## More
 
@@ -160,7 +164,7 @@ For automation or agent-driven setup, use:
 - `autotouch cli-manifest --output json` for the local machine-readable command contract
 - `autotouch cli-reference` for the shipped parser-generated reference
 - `autotouch capabilities --output json` for provider/workflow contracts
-- `autotouch --version` should be `0.2.85` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, and the cleaned single LinkedIn-sourced list-build path
+- `autotouch --version` should be `0.2.87` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, the cleaned single LinkedIn-sourced list-build path, and Exa Company Search up to 100 results
 - `autotouch capabilities --output json --select list_builds` for documented list-build inputs such as geography IDs, company size buckets, profile language, and company IDs
 - `autotouch list-build inputs` and `autotouch list-build pricing` before creating durable list-build jobs
 - `autotouch list-build companies` and `autotouch list-build leads` for durable LinkedIn-sourced company and lead list builds with Smart Table-owned background workers, visible progress, and no user-owned LinkedIn connection requirement

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/README.md

@@ -107,7 +107,7 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 - Inspect rows: `autotouch rows list`, `autotouch rows get`
 - Inspect one cell: `autotouch cells get`
 - Create a workflow column: `autotouch columns recipe`, `autotouch columns create`
-- Run neural search for niche/specific company or people discovery: `autotouch search companies`, `autotouch search people`
+- Run Exa Company Search for basic company discovery, competitors, and lookalikes: `autotouch search companies`, `autotouch search similar-companies`
 - Inspect list-build inputs: `autotouch list-build inputs`
 - Build durable company and lead lists on Smart Table workers: `autotouch list-build companies`, `autotouch list-build leads`
 - Check connected LinkedIn account status/limits: `autotouch linkedin status`, `autotouch linkedin limits`
@@ -121,13 +121,17 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 
 Autotouch has two different company/people discovery paths.
 
-Use `autotouch search companies` and `autotouch search people` for neural search when the target is niche, semantic, or hard to express with structured filters. Examples: "AI workflow platforms for healthcare", "AI workflow startups selling to law firms", or "RevOps operators at PLG SaaS companies". Neural company/people search returns at most 10 results per API call and costs 1 credit per API call.
+Use `autotouch search companies` for Exa Company Search when the target is basic company discovery, competitors, lookalikes, or a semantic company search. Put company constraints directly in natural language: industry, geography, funding, headcount, technology, founded date, and similarity. Examples: "AI workflow platforms for healthcare", "US education software startups focused on AI math tutoring with 1-50 employees", or "companies like Rippling for HR and workforce management". Exa company search returns up to 100 results per API call and costs 1 credit per 10 returned results.
 
-Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
+Use `autotouch search similar-companies` as a convenience wrapper for Exa Company Search similarity. Prefer `--query` for Exa's documented semantic patterns, such as "companies like Notion", "competitors of Stripe", or "companies working on making space travel cheaper". The older seed fields `company_name`, `company_domain`, and `company_description` remain available when you want the CLI/API to compose that query; it does not use Exa `/findSimilar`.
+
+Use `autotouch search people` for provider-hidden people search when the target is niche, semantic, or hard to express with structured filters. People search currently returns at most 10 results per API call and costs 1 credit per API call.
+
+Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable LinkedIn/Sales Navigator list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
 
 For structured company builds, start with industry IDs plus geo/company-size filters. User-supplied `keywords` are optional when structured filters are present; use them only to refine, disambiguate, or recover hard-to-classify targets instead of carrying the whole target definition.
 
-For account-first prospecting, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use neural search as a default pre-step for every list build; use it when the user's target is genuinely semantic or niche.
+For account-first prospecting from LinkedIn/Sales Navigator, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use Exa company search as a default pre-step for every LinkedIn list build; use it when the user's target is genuinely semantic, niche, competitor-based, or not meant to be LinkedIn-sourced.
 
 ## More
 
@@ -135,7 +139,7 @@ For automation or agent-driven setup, use:
 - `autotouch cli-manifest --output json` for the local machine-readable command contract
 - `autotouch cli-reference` for the shipped parser-generated reference
 - `autotouch capabilities --output json` for provider/workflow contracts
-- `autotouch --version` should be `0.2.85` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, and the cleaned single LinkedIn-sourced list-build path
+- `autotouch --version` should be `0.2.87` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, the cleaned single LinkedIn-sourced list-build path, and Exa Company Search up to 100 results
 - `autotouch capabilities --output json --select list_builds` for documented list-build inputs such as geography IDs, company size buckets, profile language, and company IDs
 - `autotouch list-build inputs` and `autotouch list-build pricing` before creating durable list-build jobs
 - `autotouch list-build companies` and `autotouch list-build leads` for durable LinkedIn-sourced company and lead list builds with Smart Table-owned background workers, visible progress, and no user-owned LinkedIn connection requirement

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/cli.py

@@ -2398,6 +2398,7 @@ cmd_auth_schema = _bound("auth_schema")
 cmd_auth_whoami = _bound("auth_whoami")
 cmd_cells_get = _bound("cells_get")
 cmd_columns_create = _bound("columns_create")
+cmd_columns_update = _bound("columns_update")
 cmd_columns_move = _bound("columns_move")
 cmd_columns_projections = _bound("columns_projections")
 cmd_columns_run_next = _bound("columns_run_next")

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/commands/columns.py

@@ -7,6 +7,16 @@ from typing import Any, Callable, Dict, Sequence
 from autotouch_cli.exceptions import AutotouchCreditError, AutotouchInputError
 
 
+_RESEARCH_MODE_ALIASES = {
+    "lite": "lite",
+    "light": "lite",
+    "quick": "lite",
+    "heavy": "heavy",
+    "deep": "heavy",
+    "thorough": "heavy",
+}
+
+
 @dataclass(frozen=True)
 class ColumnCommandRuntime:
     resolve_token: Callable[[str | None, bool], str | None]
@@ -52,6 +62,54 @@ def _apply_column_position_overrides(
     return merged
 
 
+def _normalize_research_mode_option(value: Any) -> str:
+    normalized = str(value or "").strip().lower()
+    if normalized in _RESEARCH_MODE_ALIASES:
+        return _RESEARCH_MODE_ALIASES[normalized]
+    raise AutotouchInputError("--research-mode must be one of: lite, heavy")
+
+
+def _apply_research_mode_overrides(
+    payload: Dict[str, Any],
+    args: argparse.Namespace,
+    *,
+    force_agent_on_explicit: bool,
+) -> Dict[str, Any]:
+    explicit_value = getattr(args, "research_mode", None)
+    config = payload.get("config")
+    if not isinstance(config, dict):
+        if explicit_value is None:
+            return payload
+        config = {}
+
+    current_mode = str(config.get("mode") or "").strip().lower()
+    has_research_mode = (
+        config.get("researchMode") is not None
+        or config.get("research_mode") is not None
+    )
+    should_default_agent_lite = current_mode == "agent" and not has_research_mode
+    should_clean_agent_config = current_mode == "agent" and has_research_mode
+    if explicit_value is None and not should_default_agent_lite and not should_clean_agent_config:
+        return payload
+
+    merged = dict(payload)
+    next_config = dict(config)
+    if explicit_value is not None:
+        if force_agent_on_explicit:
+            next_config["mode"] = "agent"
+        next_config["researchMode"] = _normalize_research_mode_option(explicit_value)
+    elif config.get("researchMode") is not None:
+        next_config["researchMode"] = _normalize_research_mode_option(config.get("researchMode"))
+    elif config.get("research_mode") is not None:
+        next_config["researchMode"] = _normalize_research_mode_option(config.get("research_mode"))
+    else:
+        next_config["researchMode"] = "lite"
+    next_config.pop("research_mode", None)
+    next_config.pop("lite", None)
+    merged["config"] = next_config
+    return merged
+
+
 def _normalize_http_request_test_payload(payload: Dict[str, Any]) -> Dict[str, Any]:
     candidate = payload
     config = payload.get("config")
@@ -111,6 +169,7 @@ def cmd_columns_create(args: argparse.Namespace, *, runtime: ColumnCommandRuntim
         allow_position=True,
         allow_neighbors=True,
     )
+    payload = _apply_research_mode_overrides(payload, args, force_agent_on_explicit=True)
     runtime.validate_column_contract_payload_or_exit(payload)
     runtime.validate_add_to_crm_create_payload(payload)
     data = runtime.request_api(
@@ -156,6 +215,7 @@ def cmd_columns_update(args: argparse.Namespace, *, runtime: ColumnCommandRuntim
         allow_position=True,
         allow_neighbors=False,
     )
+    payload = _apply_research_mode_overrides(payload, args, force_agent_on_explicit=False)
     runtime.validate_column_contract_payload_or_exit(payload)
     data = runtime.request_api(
         "PATCH",
@@ -401,6 +461,11 @@ def register_columns_subcommands(
     pcc.add_argument("--position", type=float, help="Explicit numeric position override")
     pcc.add_argument("--after-column-id", help="Place the new column after this existing column id")
     pcc.add_argument("--before-column-id", help="Place the new column before this existing column id")
+    pcc.add_argument(
+        "--research-mode",
+        choices=["lite", "heavy"],
+        help="For LLM Research Agent columns, set config.researchMode. Omitted agent columns default to lite.",
+    )
     add_api_common_arguments(pcc)
     pcc.set_defaults(func=handlers["create"])
 
@@ -417,6 +482,11 @@ def register_columns_subcommands(
     pcu.add_argument("--data-json", help="ColumnUpdate payload JSON")
     pcu.add_argument("--data-file", help="ColumnUpdate payload file path")
     pcu.add_argument("--position", type=float, help="Explicit numeric position override")
+    pcu.add_argument(
+        "--research-mode",
+        choices=["lite", "heavy"],
+        help="Set config.researchMode. Include config.mode=agent when switching a Quick LLM column to Research Agent.",
+    )
     add_api_common_arguments(pcu)
     pcu.set_defaults(func=handlers["update"])
 
@@ -545,6 +615,12 @@ def register_columns_subcommands(
 
     pcre = col_sub.add_parser("recipe", help="Print copy-ready ColumnCreate payload templates")
     pcre.add_argument("--type", choices=["all", *column_recipe_types], default="all")
+    pcre.add_argument(
+        "--research-mode",
+        choices=["lite", "heavy"],
+        default="lite",
+        help="Research Agent depth for LLM recipes (default: lite; heavy costs more per row)",
+    )
     pcre.add_argument("--out-file", help="Write template JSON to file (single type writes payload only)")
     add_output_formatting_arguments(pcre)
     pcre.set_defaults(func=handlers["recipe"])

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/commands/search.py

@@ -67,7 +67,7 @@ def cmd_search_companies(args: argparse.Namespace, *, runtime: SearchCommandRunt
         runtime=runtime,
         default_payload={
             "query": str(getattr(args, "query", "") or "").strip(),
-            "limit": max(1, min(int(getattr(args, "limit", 10) or 10), 10)),
+            "limit": max(1, min(int(getattr(args, "limit", 10) or 10), 100)),
         },
     )
     data = runtime.request_api(
@@ -120,11 +120,16 @@ def cmd_search_people(args: argparse.Namespace, *, runtime: SearchCommandRuntime
 def cmd_search_similar_companies(args: argparse.Namespace, *, runtime: SearchCommandRuntime) -> None:
     token = runtime.resolve_token(args.token, required=True)
     payload: Dict[str, Any] = {
-        "company_name": str(getattr(args, "company_name", "") or "").strip(),
-        "limit": max(1, min(int(getattr(args, "limit", 10) or 10), 10)),
+        "limit": max(1, min(int(getattr(args, "limit", 10) or 10), 100)),
     }
+    query = str(getattr(args, "query", "") or "").strip()
+    company_name = str(getattr(args, "company_name", "") or "").strip()
     company_domain = str(getattr(args, "company_domain", "") or "").strip()
     company_description = str(getattr(args, "company_description", "") or "").strip()
+    if query:
+        payload["query"] = query
+    if company_name:
+        payload["company_name"] = company_name
     if company_domain:
         payload["company_domain"] = company_domain
     if company_description:
@@ -264,9 +269,9 @@ def register_search_subcommands(
     add_api_common_arguments(recipe_parser)
     recipe_parser.set_defaults(func=handlers["recipe"])
 
-    companies_parser = search_sub.add_parser("companies", help="Find companies with provider-hidden neural search")
-    companies_parser.add_argument("--query", help="Free-text company query")
-    companies_parser.add_argument("--limit", type=int, default=10, help="Max results, 1-10 (default: 10)")
+    companies_parser = search_sub.add_parser("companies", help="Find companies with Exa Company Search")
+    companies_parser.add_argument("--query", help="Natural-language company query")
+    companies_parser.add_argument("--limit", type=int, default=10, help="Max results, 1-100 (default: 10)")
     companies_parser.add_argument("--data-json", help="Explicit request payload JSON")
     companies_parser.add_argument("--data-file", help="Path to request payload JSON file")
     add_api_common_arguments(companies_parser)
@@ -284,10 +289,11 @@ def register_search_subcommands(
     people_parser.set_defaults(func=handlers["people"])
 
     similar_parser = search_sub.add_parser("similar-companies", help="Find companies similar to a seed company")
+    similar_parser.add_argument("--query", help="Exa Company Search semantic similarity query, e.g. 'companies like Notion for enterprise knowledge management'")
     similar_parser.add_argument("--company-name", help="Seed company name")
     similar_parser.add_argument("--company-domain", help="Official company domain to ground similarity search")
     similar_parser.add_argument("--company-description", help="Short semantic company description to improve peer matching")
-    similar_parser.add_argument("--limit", type=int, default=10, help="Max results, 1-10 (default: 10)")
+    similar_parser.add_argument("--limit", type=int, default=10, help="Max results, 1-100 (default: 10)")
     similar_parser.add_argument("--data-json", help="Explicit request payload JSON")
     similar_parser.add_argument("--data-file", help="Path to request payload JSON file")
     add_api_common_arguments(similar_parser)

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/data/CLI_REFERENCE.md

@@ -1,6 +1,6 @@
 # Autotouch CLI Reference
 
-Generated from the installed parser for `autotouch-cli` `0.2.85`.
+Generated from the installed parser for `autotouch-cli` `0.2.87`.
 Manifest schema version: `2`.
 
 ## Output Modes
@@ -1079,6 +1079,7 @@ For llm_enrichment, there are two valid authoring paths:
                                 [--data-file DATA_FILE] [--position POSITION]
                                 [--after-column-id AFTER_COLUMN_ID]
                                 [--before-column-id BEFORE_COLUMN_ID]
+                                [--research-mode {lite,heavy}]
                                 [--base-url BASE_URL] [--token TOKEN]
                                 [--use-x-api-key] [--timeout TIMEOUT]
                                 [--output {json,ndjson,human}] [--compact]
@@ -1100,6 +1101,7 @@ The runnable prompt is always config.advancedPrompt before execution.
   - `--position` (kind=number): Explicit numeric position override
   - `--after-column-id` (kind=string): Place the new column after this existing column id
   - `--before-column-id` (kind=string): Place the new column before this existing column id
+  - `--research-mode` (kind=string; choices=lite,heavy): For LLM Research Agent columns, set config.researchMode. Omitted agent columns default to lite.
   - `--base-url` (kind=string; default=https://app.autotouch.ai): API base URL (default: https://app.autotouch.ai)
   - `--token` (kind=string; sensitive): Developer API key / JWT token
   - `--use-x-api-key` (kind=boolean; when omitted=False; when present=True): Send token via X-API-Key header
@@ -1320,13 +1322,15 @@ Print copy-ready ColumnCreate payload templates
 - Output modes: `json, ndjson, human`
 - Example:
   - `autotouch columns recipe [-h]
-                                [--type {all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads}]
+                                [--type {all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads,sync_to_leads_array,add_to_leads_array}]
+                                [--research-mode {lite,heavy}]
                                 [--out-file OUT_FILE]
                                 [--output {json,ndjson,human}] [--compact]
                                 [--select SELECT |
                                 --json-pointer JSON_POINTER]`
 - Options:
-  - `--type` (kind=string; choices=all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads; default=all)
+  - `--type` (kind=string; choices=all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads,sync_to_leads_array,add_to_leads_array; default=all)
+  - `--research-mode` (kind=string; choices=lite,heavy; default=lite): Research Agent depth for LLM recipes (default: lite; heavy costs more per row)
   - `--out-file` (kind=file; input=file): Write template JSON to file (single type writes payload only)
   - `--output` (kind=string; choices=json,ndjson,human; default=json): Output mode
   - `--compact` (kind=boolean; when omitted=False; when present=True): Print compact JSON
@@ -1538,6 +1542,7 @@ Update a column definition. Nested config keys merge into the existing config, s
   - `autotouch columns update [-h] --table-id TABLE_ID --column-id COLUMN_ID
                                 [--data-json DATA_JSON]
                                 [--data-file DATA_FILE] [--position POSITION]
+                                [--research-mode {lite,heavy}]
                                 [--base-url BASE_URL] [--token TOKEN]
                                 [--use-x-api-key] [--timeout TIMEOUT]
                                 [--output {json,ndjson,human}] [--compact]
@@ -1549,6 +1554,7 @@ Update a column definition. Nested config keys merge into the existing config, s
   - `--data-json` (kind=json; input=json): ColumnUpdate payload JSON
   - `--data-file` (kind=file; input=file): ColumnUpdate payload file path
   - `--position` (kind=number): Explicit numeric position override
+  - `--research-mode` (kind=string; choices=lite,heavy): Set config.researchMode. Include config.mode=agent when switching a Quick LLM column to Research Agent.
   - `--base-url` (kind=string; default=https://app.autotouch.ai): API base URL (default: https://app.autotouch.ai)
   - `--token` (kind=string; sensitive): Developer API key / JWT token
   - `--use-x-api-key` (kind=boolean; when omitted=False; when present=True): Send token via X-API-Key header
@@ -3553,7 +3559,7 @@ Provider-hidden search operations
 
 #### `autotouch search companies`
 
-Find companies with provider-hidden neural search
+Find companies with Exa Company Search
 
 - Auth: `developer_key_or_user_session`
 - Stability: `stable`
@@ -3569,8 +3575,8 @@ Find companies with provider-hidden neural search
                                   [--select SELECT |
                                   --json-pointer JSON_POINTER] [--verbose]`
 - Options:
-  - `--query` (kind=string): Free-text company query
-  - `--limit` (kind=integer; default=10): Max results, 1-10 (default: 10)
+  - `--query` (kind=string): Natural-language company query
+  - `--limit` (kind=integer; default=10): Max results, 1-100 (default: 10)
   - `--data-json` (kind=json; input=json): Explicit request payload JSON
   - `--data-file` (kind=file; input=file): Path to request payload JSON file
   - `--base-url` (kind=string; default=https://app.autotouch.ai): API base URL (default: https://app.autotouch.ai)
@@ -3805,7 +3811,8 @@ Find companies similar to a seed company
 - Destructive: `no`
 - Output modes: `json, ndjson, human`
 - Example:
-  - `autotouch search similar-companies [-h] [--company-name COMPANY_NAME]
+  - `autotouch search similar-companies [-h] [--query QUERY]
+                                          [--company-name COMPANY_NAME]
                                           [--company-domain COMPANY_DOMAIN]
                                           [--company-description COMPANY_DESCRIPTION]
                                           [--limit LIMIT]
@@ -3819,10 +3826,11 @@ Find companies similar to a seed company
                                           --json-pointer JSON_POINTER]
                                           [--verbose]`
 - Options:
+  - `--query` (kind=string): Exa Company Search semantic similarity query, e.g. 'companies like Notion for enterprise knowledge management'
   - `--company-name` (kind=string): Seed company name
   - `--company-domain` (kind=string): Official company domain to ground similarity search
   - `--company-description` (kind=string): Short semantic company description to improve peer matching
-  - `--limit` (kind=integer; default=10): Max results, 1-10 (default: 10)
+  - `--limit` (kind=integer; default=10): Max results, 1-100 (default: 10)
   - `--data-json` (kind=json; input=json): Explicit request payload JSON
   - `--data-file` (kind=file; input=file): Path to request payload JSON file
   - `--base-url` (kind=string; default=https://app.autotouch.ai): API base URL (default: https://app.autotouch.ai)

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/data/cli-manifest.json

@@ -1,5 +1,5 @@
 {
-  "version": "0.2.85",
+  "version": "0.2.87",
   "manifest_schema_version": 2,
   "entry_points": {
     "autotouch": "autotouch_cli.cli:main",
@@ -6304,7 +6304,7 @@
       ],
       "aliases": [],
       "group": "search",
-      "help": "Find companies with provider-hidden neural search",
+      "help": "Find companies with Exa Company Search",
       "description": null,
       "notes": null,
       "required_flags": [],
@@ -6313,7 +6313,7 @@
         {
           "dest": "query",
           "required": false,
-          "help": "Free-text company query",
+          "help": "Natural-language company query",
           "kind": "string",
           "action": "store",
           "flags": [
@@ -6324,7 +6324,7 @@
         {
           "dest": "limit",
           "required": false,
-          "help": "Max results, 1-10 (default: 10)",
+          "help": "Max results, 1-100 (default: 10)",
           "kind": "integer",
           "action": "store",
           "default_when_omitted": 10,
@@ -6785,6 +6785,17 @@
       "required_flags": [],
       "positionals": [],
       "options": [
+        {
+          "dest": "query",
+          "required": false,
+          "help": "Exa Company Search semantic similarity query, e.g. 'companies like Notion for enterprise knowledge management'",
+          "kind": "string",
+          "action": "store",
+          "flags": [
+            "--query"
+          ],
+          "takes_value": true
+        },
         {
           "dest": "company_name",
           "required": false,
@@ -6821,7 +6832,7 @@
         {
           "dest": "limit",
           "required": false,
-          "help": "Max results, 1-10 (default: 10)",
+          "help": "Max results, 1-100 (default: 10)",
           "kind": "integer",
           "action": "store",
           "default_when_omitted": 10,
@@ -7008,7 +7019,7 @@
       },
       "availability": null,
       "examples": [
-        "autotouch search similar-companies [-h] [--company-name COMPANY_NAME]\n                                          [--company-domain COMPANY_DOMAIN]\n                                          [--company-description COMPANY_DESCRIPTION]\n                                          [--limit LIMIT]\n                                          [--data-json DATA_JSON]\n                                          [--data-file DATA_FILE]\n                                          [--base-url BASE_URL]\n                                          [--token TOKEN] [--use-x-api-key]\n                                          [--timeout TIMEOUT]\n                                          [--output {json,ndjson,human}]\n                                          [--compact] [--select SELECT |\n                                          --json-pointer JSON_POINTER]\n                                          [--verbose]"
+        "autotouch search similar-companies [-h] [--query QUERY]\n                                          [--company-name COMPANY_NAME]\n                                          [--company-domain COMPANY_DOMAIN]\n                                          [--company-description COMPANY_DESCRIPTION]\n                                          [--limit LIMIT]\n                                          [--data-json DATA_JSON]\n                                          [--data-file DATA_FILE]\n                                          [--base-url BASE_URL]\n                                          [--token TOKEN] [--use-x-api-key]\n                                          [--timeout TIMEOUT]\n                                          [--output {json,ndjson,human}]\n                                          [--compact] [--select SELECT |\n                                          --json-pointer JSON_POINTER]\n                                          [--verbose]"
       ],
       "stability": "stable",
       "handler": "_fn"
@@ -14852,6 +14863,21 @@
           ],
           "takes_value": true
         },
+        {
+          "dest": "research_mode",
+          "required": false,
+          "help": "For LLM Research Agent columns, set config.researchMode. Omitted agent columns default to lite.",
+          "kind": "string",
+          "action": "store",
+          "choices": [
+            "lite",
+            "heavy"
+          ],
+          "flags": [
+            "--research-mode"
+          ],
+          "takes_value": true
+        },
         {
           "dest": "base_url",
           "required": false,
@@ -15005,7 +15031,7 @@
       },
       "availability": null,
       "examples": [
-        "autotouch columns create [-h] --table-id TABLE_ID\n                                [--data-json DATA_JSON]\n                                [--data-file DATA_FILE] [--position POSITION]\n                                [--after-column-id AFTER_COLUMN_ID]\n                                [--before-column-id BEFORE_COLUMN_ID]\n                                [--base-url BASE_URL] [--token TOKEN]\n                                [--use-x-api-key] [--timeout TIMEOUT]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER] [--verbose]"
+        "autotouch columns create [-h] --table-id TABLE_ID\n                                [--data-json DATA_JSON]\n                                [--data-file DATA_FILE] [--position POSITION]\n                                [--after-column-id AFTER_COLUMN_ID]\n                                [--before-column-id BEFORE_COLUMN_ID]\n                                [--research-mode {lite,heavy}]\n                                [--base-url BASE_URL] [--token TOKEN]\n                                [--use-x-api-key] [--timeout TIMEOUT]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER] [--verbose]"
       ],
       "stability": "stable",
       "handler": "_fn"
@@ -15084,6 +15110,21 @@
           ],
           "takes_value": true
         },
+        {
+          "dest": "research_mode",
+          "required": false,
+          "help": "Set config.researchMode. Include config.mode=agent when switching a Quick LLM column to Research Agent.",
+          "kind": "string",
+          "action": "store",
+          "choices": [
+            "lite",
+            "heavy"
+          ],
+          "flags": [
+            "--research-mode"
+          ],
+          "takes_value": true
+        },
         {
           "dest": "base_url",
           "required": false,
@@ -15237,7 +15278,7 @@
       },
       "availability": null,
       "examples": [
-        "autotouch columns update [-h] --table-id TABLE_ID --column-id COLUMN_ID\n                                [--data-json DATA_JSON]\n                                [--data-file DATA_FILE] [--position POSITION]\n                                [--base-url BASE_URL] [--token TOKEN]\n                                [--use-x-api-key] [--timeout TIMEOUT]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER] [--verbose]"
+        "autotouch columns update [-h] --table-id TABLE_ID --column-id COLUMN_ID\n                                [--data-json DATA_JSON]\n                                [--data-file DATA_FILE] [--position POSITION]\n                                [--research-mode {lite,heavy}]\n                                [--base-url BASE_URL] [--token TOKEN]\n                                [--use-x-api-key] [--timeout TIMEOUT]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER] [--verbose]"
       ],
       "stability": "stable",
       "handler": "_fn"
@@ -17555,7 +17596,9 @@
             "add_to_sequence",
             "http_request",
             "add_to_leads",
-            "sync_to_leads"
+            "sync_to_leads",
+            "sync_to_leads_array",
+            "add_to_leads_array"
           ],
           "default_when_omitted": "all",
           "default": "all",
@@ -17564,6 +17607,23 @@
           ],
           "takes_value": true
         },
+        {
+          "dest": "research_mode",
+          "required": false,
+          "help": "Research Agent depth for LLM recipes (default: lite; heavy costs more per row)",
+          "kind": "string",
+          "action": "store",
+          "choices": [
+            "lite",
+            "heavy"
+          ],
+          "default_when_omitted": "lite",
+          "default": "lite",
+          "flags": [
+            "--research-mode"
+          ],
+          "takes_value": true
+        },
         {
           "dest": "out_file",
           "required": false,
@@ -17663,7 +17723,7 @@
       },
       "availability": null,
       "examples": [
-        "autotouch columns recipe [-h]\n                                [--type {all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads}]\n                                [--out-file OUT_FILE]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER]"
+        "autotouch columns recipe [-h]\n                                [--type {all,formatter,llm_enrichment,email_finder,phone_finder,lead_finder,add_to_crm,sync_to_table,add_to_sequence,http_request,add_to_leads,sync_to_leads,sync_to_leads_array,add_to_leads_array}]\n                                [--research-mode {lite,heavy}]\n                                [--out-file OUT_FILE]\n                                [--output {json,ndjson,human}] [--compact]\n                                [--select SELECT |\n                                --json-pointer JSON_POINTER]"
       ],
       "stability": "stable",
       "handler": "_fn"

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli/templates.py

@@ -51,6 +51,69 @@ COLUMN_RECIPE_ALIASES: Dict[str, Dict[str, Any]] = {
             "Emits the standard Sync to Leads payload. Internal provider/config names remain add_to_crm.",
         ],
     },
+    "sync_to_leads_array": {
+        "canonical": "add_to_crm",
+        "key": "sync_contact_list_to_leads",
+        "label": "Sync Contact List to Leads",
+        "payload": {
+            "key": "sync_contact_list_to_leads",
+            "label": "Sync Contact List to Leads",
+            "kind": "enrichment",
+            "dataType": "json",
+            "origin": "manual",
+            "autoRun": "onSourceUpdate",
+            "config": {
+                "provider": "add_to_crm",
+                "requireCompanyDomain": True,
+                "leadSource": "research_table_export",
+                "fieldMappings": {
+                    "mode": "array",
+                    "sourceColumn": "<CONTACT_LIST_COLUMN_KEY>",
+                    "sourceArrayPath": "contacts",
+                    "common": {
+                        "companyDomain": "domain",
+                        "companyName": "company_name",
+                    },
+                    "template": {
+                        "linkedinUrl": "linkedin_url",
+                        "firstName": "first_name",
+                        "lastName": "last_name",
+                        "title": "title",
+                        "companyName": "company_name",
+                        "companyDomain": "company_domain",
+                        "emailAddresses": [
+                            {"column": "work_email", "type": "work"},
+                            {"column": "email", "type": "work"},
+                        ],
+                        "phoneNumbers": [
+                            {"column": "mobile_phone", "type": "mobile"},
+                            {"column": "phone", "type": "mobile"},
+                        ],
+                    },
+                },
+                "sourceColumns": [
+                    "<CONTACT_LIST_COLUMN_KEY>",
+                    "domain",
+                    "company_name",
+                ],
+            },
+        },
+        "notes": [
+            "Use this when an upstream research column returns a contact/person array such as contacts[], people[], leads[], or items[] and the goal is to write Leads directly.",
+            "Set config.fieldMappings.sourceColumn to the upstream column key, for example key_buyer_leads.",
+            "Set config.fieldMappings.sourceArrayPath to the array path inside that JSON value; use contacts for key_buyer_leads.contacts[] or items for canonical items[].",
+            "Template fields are item-relative: linkedinUrl=linkedin_url reads each array item's linkedin_url.",
+            "Common fields are parent-row mappings used for every item, usually domain/company_name.",
+            "Prefer this over manual row extraction, local JSON files, and rows add/import when writing discovered contacts into Leads CRM.",
+        ],
+    },
+    "add_to_leads_array": {
+        "canonical": "add_to_crm",
+        "alias_of": "sync_to_leads_array",
+        "notes": [
+            "Legacy-friendly alias for sync_to_leads_array.",
+        ],
+    },
 }
 COLUMN_RECIPE_TYPES = tuple(dict.fromkeys([*COLUMN_RECIPE_BASE_TYPES, *COLUMN_RECIPE_ALIASES.keys()]))
 WORKFLOW_BLUEPRINT_TYPES = tuple(sorted(automation_workflow_blueprints().keys()))
@@ -103,11 +166,16 @@ def _build_column_recipe_catalog() -> Tuple[Dict[str, Dict[str, Any]], Dict[str,
             recipes[recipe_type] = payload
         notes[recipe_type] = contract.recipe_notes_list()
     for alias, metadata in COLUMN_RECIPE_ALIASES.items():
+        alias_of = str(metadata.get("alias_of") or "").strip()
+        if alias_of and alias_of in recipes:
+            recipes[alias] = json.loads(json.dumps(recipes[alias_of]))
+            notes[alias] = list(metadata.get("notes") or []) + list(notes.get(alias_of) or [])
+            continue
         canonical = str(metadata.get("canonical") or "").strip()
         canonical_payload = recipes.get(canonical)
         if not isinstance(canonical_payload, dict):
             continue
-        alias_payload = json.loads(json.dumps(canonical_payload))
+        alias_payload = json.loads(json.dumps(metadata.get("payload") or canonical_payload))
         if metadata.get("key"):
             alias_payload["key"] = metadata["key"]
         if metadata.get("label"):
@@ -917,6 +985,32 @@ def _recipe_payload(recipe_type: str) -> Dict[str, Any]:
     return _json_copy(payload)
 
 
+def _apply_recipe_research_mode(payload: Dict[str, Any], research_mode: Any) -> Dict[str, Any]:
+    config = payload.get("config")
+    if not isinstance(config, dict):
+        return payload
+    if str(config.get("mode") or "").strip().lower() != "agent":
+        return payload
+    mode = str(research_mode or "lite").strip().lower()
+    if mode not in {"lite", "heavy"}:
+        mode = "lite"
+    payload = dict(payload)
+    payload["config"] = {
+        **config,
+        "researchMode": mode,
+    }
+    payload["config"].pop("research_mode", None)
+    payload["config"].pop("lite", None)
+    return payload
+
+
+def _recipe_payload_for_args(recipe_type: str, args: argparse.Namespace) -> Dict[str, Any]:
+    return _apply_recipe_research_mode(
+        _recipe_payload(recipe_type),
+        getattr(args, "research_mode", "lite"),
+    )
+
+
 def _sequence_recipe_payload(recipe_type: str) -> Dict[str, Any]:
     payload = SEQUENCE_RECIPES.get(recipe_type)
     if not payload:
@@ -1639,7 +1733,7 @@ def emit_workflows_scaffold(args: argparse.Namespace, runtime: TemplateRuntime)
 def emit_columns_recipe(args: argparse.Namespace, runtime: TemplateRuntime) -> None:
     selected_type = str(args.type or "all")
     if selected_type == "all":
-        recipes = {name: _recipe_payload(name) for name in COLUMN_RECIPE_TYPES}
+        recipes = {name: _recipe_payload_for_args(name, args) for name in COLUMN_RECIPE_TYPES}
         output = {
             "recipes": recipes,
             "notes": COLUMN_RECIPE_NOTES,
@@ -1662,7 +1756,7 @@ def emit_columns_recipe(args: argparse.Namespace, runtime: TemplateRuntime) -> N
         runtime.print_json(output, getattr(args, "compact", False))
         return
 
-    payload = _recipe_payload(selected_type)
+    payload = _recipe_payload_for_args(selected_type, args)
     output_single = {
         "type": selected_type,
         "payload": payload,

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_cli.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: autotouch-cli
-Version: 0.2.85
+Version: 0.2.87
 Summary: Autotouch Smart Table CLI
 Project-URL: Homepage, https://app.autotouch.ai
 Project-URL: Documentation, https://github.com/nicolonic/autotouch_main/tree/main/docs/research-table/reference
@@ -132,7 +132,7 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 - Inspect rows: `autotouch rows list`, `autotouch rows get`
 - Inspect one cell: `autotouch cells get`
 - Create a workflow column: `autotouch columns recipe`, `autotouch columns create`
-- Run neural search for niche/specific company or people discovery: `autotouch search companies`, `autotouch search people`
+- Run Exa Company Search for basic company discovery, competitors, and lookalikes: `autotouch search companies`, `autotouch search similar-companies`
 - Inspect list-build inputs: `autotouch list-build inputs`
 - Build durable company and lead lists on Smart Table workers: `autotouch list-build companies`, `autotouch list-build leads`
 - Check connected LinkedIn account status/limits: `autotouch linkedin status`, `autotouch linkedin limits`
@@ -146,13 +146,17 @@ autotouch rows get --table-id "$TABLE_ID" --row-id "$ROW_ID" --output json
 
 Autotouch has two different company/people discovery paths.
 
-Use `autotouch search companies` and `autotouch search people` for neural search when the target is niche, semantic, or hard to express with structured filters. Examples: "AI workflow platforms for healthcare", "AI workflow startups selling to law firms", or "RevOps operators at PLG SaaS companies". Neural company/people search returns at most 10 results per API call and costs 1 credit per API call.
+Use `autotouch search companies` for Exa Company Search when the target is basic company discovery, competitors, lookalikes, or a semantic company search. Put company constraints directly in natural language: industry, geography, funding, headcount, technology, founded date, and similarity. Examples: "AI workflow platforms for healthcare", "US education software startups focused on AI math tutoring with 1-50 employees", or "companies like Rippling for HR and workforce management". Exa company search returns up to 100 results per API call and costs 1 credit per 10 returned results.
 
-Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
+Use `autotouch search similar-companies` as a convenience wrapper for Exa Company Search similarity. Prefer `--query` for Exa's documented semantic patterns, such as "companies like Notion", "competitors of Stripe", or "companies working on making space travel cheaper". The older seed fields `company_name`, `company_domain`, and `company_description` remain available when you want the CLI/API to compose that query; it does not use Exa `/findSimilar`.
+
+Use `autotouch search people` for provider-hidden people search when the target is niche, semantic, or hard to express with structured filters. People search currently returns at most 10 results per API call and costs 1 credit per API call.
+
+Use `autotouch list-build companies` and `autotouch list-build leads` for structured, repeatable LinkedIn/Sales Navigator list construction. This is the only CLI path for building LinkedIn-sourced company and lead lists. Treat list builds as the sourcing step for the Smart Table research workspace: build company/account records first by default, add the records to a research table, then enrich, score, segment, find related leads, attach signals/notes, and continue downstream workflows from that table. Use lead builds directly when the request is explicitly person/contact focused. This path uses Autotouch-managed provider access, so users do not need to connect their own LinkedIn account. Use filters like geography IDs, company size, industry IDs, title/persona, and current company IDs; call `autotouch list-build inputs` for supported values. Durable list builds run as background jobs with status/results endpoints and cost 1 credit per successful non-empty result page.
 
 For structured company builds, start with industry IDs plus geo/company-size filters. User-supplied `keywords` are optional when structured filters are present; use them only to refine, disambiguate, or recover hard-to-classify targets instead of carrying the whole target definition.
 
-For account-first prospecting, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use neural search as a default pre-step for every list build; use it when the user's target is genuinely semantic or niche.
+For account-first prospecting from LinkedIn/Sales Navigator, build companies/accounts first, inspect the returned company IDs, then pass those IDs to `autotouch list-build leads --current-company-id ...`. Do not use Exa company search as a default pre-step for every LinkedIn list build; use it when the user's target is genuinely semantic, niche, competitor-based, or not meant to be LinkedIn-sourced.
 
 ## More
 
@@ -160,7 +164,7 @@ For automation or agent-driven setup, use:
 - `autotouch cli-manifest --output json` for the local machine-readable command contract
 - `autotouch cli-reference` for the shipped parser-generated reference
 - `autotouch capabilities --output json` for provider/workflow contracts
-- `autotouch --version` should be `0.2.85` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, and the cleaned single LinkedIn-sourced list-build path
+- `autotouch --version` should be `0.2.87` or newer for structured company list builds without user-supplied keywords, stored list-build input lookup, research-workspace list-build guidance, the cleaned single LinkedIn-sourced list-build path, and Exa Company Search up to 100 results
 - `autotouch capabilities --output json --select list_builds` for documented list-build inputs such as geography IDs, company size buckets, profile language, and company IDs
 - `autotouch list-build inputs` and `autotouch list-build pricing` before creating durable list-build jobs
 - `autotouch list-build companies` and `autotouch list-build leads` for durable LinkedIn-sourced company and lead list builds with Smart Table-owned background workers, visible progress, and no user-owned LinkedIn connection requirement

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_shared/linkedin_contract.py

@@ -63,6 +63,8 @@ _LINKEDIN_CAPABILITIES: Dict[str, Any] = {
         "autotouch linkedin status",
         "autotouch linkedin limits",
         "autotouch list-build inputs",
+        "autotouch list-build inputs --type geo --query 'United States'",
+        "autotouch list-build inputs --type industry --query 'software'",
         "autotouch list-build companies --geo-id 103644278 --industry-id <industry_id> --company-size 1-10 --num-results 100 --wait",
     ],
 }

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_shared/provider_registry.py

@@ -378,6 +378,7 @@ _PROVIDER_CONTRACTS: Tuple[ResearchTableProviderContract, ...] = (
             "config": {
                 "instructions": "Research the company in this row and return JSON with icp_fit, summary, and risks.",
                 "mode": "agent",
+                "researchMode": "lite",
                 "temperature": 0.7,
                 "promptSource": "generated",
                 "useCompanyContext": True,
@@ -397,6 +398,7 @@ _PROVIDER_CONTRACTS: Tuple[ResearchTableProviderContract, ...] = (
             "Runtime only injects values for placeholders the prompt explicitly references.",
             "Manual/basic prompts should use explicit placeholders like {{company_name}}, {{domain}}, or {{user_value_proposition}} when those values are needed.",
             "Agent mode is JSON-oriented.",
+            "Agent research depth lives in config.researchMode and defaults to lite; set heavy only when the broader 2-credit research pass is needed.",
             "Basic mode can be text or JSON; JSON requires dataType=json.",
         ),
         column_type={
@@ -835,16 +837,47 @@ _PROVIDER_CONTRACTS: Tuple[ResearchTableProviderContract, ...] = (
                     {"linkedinUrl": {"column": "linkedin_lookup", "path": "linkedin_url"}},
                 ],
             },
+            "array_mode": {
+                "mode": "config.fieldMappings.mode=array",
+                "source_column_field": "config.fieldMappings.sourceColumn",
+                "array_path_field": "config.fieldMappings.sourceArrayPath",
+                "default_array_path": "items",
+                "common_field": "config.fieldMappings.common",
+                "template_field": "config.fieldMappings.template",
+                "template_scope": "item-relative",
+                "recommended_recipe": "autotouch columns recipe --type sync_to_leads_array",
+                "example": {
+                    "sourceColumn": "key_buyer_leads",
+                    "sourceArrayPath": "contacts",
+                    "template": {
+                        "firstName": "first_name",
+                        "lastName": "last_name",
+                        "linkedinUrl": "linkedin_url",
+                    },
+                },
+            },
         },
         composition_contract={
             "consumes": {
                 "kind": "crm_export_inputs",
-                "shape": "mapped source row values including company domain plus hard identity or, in single mode, provisional firstName/lastName fallback; structured source columns use explicit {column, path} mappings to identify the exact field",
+                "shape": "mapped source row values, multi-slot mappings, or array-mode contact/person items from a structured source column; strict mode requires company domain plus hard identity or, in single mode, provisional firstName/lastName fallback",
             },
             "produces": {
                 "kind": "lead_ids",
                 "shape": "lead creation/update result with CRM lead ids stored on the source row",
             },
+            "decision_hints": {
+                "prefer_when": [
+                    "the target output is Leads CRM, not a separate contacts research table",
+                    "an upstream research column already returns contacts[], people[], leads[], or items[]",
+                ],
+                "prefer_sync_to_table_when": [
+                    "you need a destination contacts table for per-contact enrichment, review, export, or sequencing context before CRM sync",
+                ],
+                "avoid": [
+                    "manual rows list -> local JSON transform -> rows add/import for contacts produced by an upstream column",
+                ],
+            },
             "common_next_steps": [
                 "add_to_sequence using the column key that stores those lead IDs",
             ],

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/autotouch_shared/search_contract.py

@@ -1,4 +1,4 @@
-"""Shared provider-hidden search contract for API capabilities, CLI recipes, and docs."""
+"""Shared search contract for Exa company search, people search, and related API capabilities."""
 
 from __future__ import annotations
 
@@ -15,8 +15,8 @@ _SEARCH_ENDPOINTS: Dict[str, Dict[str, Any]] = {
         "path": "/api/search/companies",
         "required_scope": SEARCH_SCOPE,
         "body": {
-            "query": "required free-text company search query",
-            "limit": "optional 1-10 (default 10)",
+            "query": "required Exa Company Search natural-language query; include industry, geography, funding, headcount, technology, and similarity constraints in the query",
+            "limit": "optional 1-100 (default 10)",
         },
         "response": {"items_field": "items", "item_type": "company"},
     },
@@ -38,10 +38,11 @@ _SEARCH_ENDPOINTS: Dict[str, Dict[str, Any]] = {
         "path": "/api/search/similar-companies",
         "required_scope": SEARCH_SCOPE,
         "body": {
-            "company_name": "required seed company name",
-            "company_domain": "optional official company domain",
-            "company_description": "optional short semantic description to improve peer matching",
-            "limit": "optional 1-10 (default 10)",
+            "query": "optional Exa Company Search semantic similarity query; use this for direct Exa query patterns like 'companies like Notion' or 'competitors of Stripe'",
+            "company_name": "optional seed company name used when query is omitted",
+            "company_domain": "optional official company domain used to disambiguate the seed in the Exa Company Search query",
+            "company_description": "optional semantic description appended to the Exa Company Search query",
+            "limit": "optional 1-100 (default 10)",
         },
         "response": {"items_field": "items", "item_type": "company"},
     },
@@ -108,13 +109,14 @@ _SEARCH_CAPABILITIES: Dict[str, Any] = {
         "use_company_search_when": [
             "you need to discover target accounts before looking for contacts",
             "the search criteria are market, product, or industry driven",
+            "the user wants basic company discovery, competitors, lookalikes, or a semantic company search",
         ],
         "use_people_search_when": [
             "you already know the target companies",
             "the search criteria are role, persona, or skill driven",
         ],
         "use_list_build_instead_when": [
-            "the user wants a LinkedIn-sourced company or lead list",
+            "the user wants a LinkedIn/Sales Navigator-sourced company or lead list",
             "the search can be expressed with geography, industry, company size, title, or current-company filters",
             "the desired output is a durable background job with status/results",
         ],
@@ -122,9 +124,9 @@ _SEARCH_CAPABILITIES: Dict[str, Any] = {
     "pricing": {
         "charge_on": "api_call_or_actual_results_returned",
         "families": {
-            "companies": {"credits_per_call": 1, "max_results_per_call": 10, "rule": "1 credit per neural search API call"},
+            "companies": {"credits_per_bucket": 1, "bucket_size": 10, "max_results_per_call": 100, "rule": "1 credit per 10 returned Exa company results"},
             "people": {"credits_per_call": 1, "max_results_per_call": 10, "rule": "1 credit per neural search API call"},
-            "similar_companies": {"credits_per_call": 1, "max_results_per_call": 10, "rule": "1 credit per neural search API call"},
+            "similar_companies": {"credits_per_bucket": 1, "bucket_size": 10, "max_results_per_call": 100, "rule": "1 credit per 10 returned Exa company results"},
             "news": {"credits_per_bucket": 1, "bucket_size": 25, "rule": "1 credit per 25 results"},
             "local_businesses": {"credits_per_bucket": 1, "bucket_size": 25, "rule": "1 credit per 25 results"},
             "reviews": {"credits_per_bucket": 1, "bucket_size": 25, "rule": "1 credit per 25 results"},
@@ -132,8 +134,9 @@ _SEARCH_CAPABILITIES: Dict[str, Any] = {
         },
     },
     "recommended_cli": [
-        "autotouch search companies --query 'remote IT asset management' --limit 25",
-        "autotouch search people --query 'VP IT' --company 'Rippling' --limit 25",
+        "autotouch search companies --query 'remote IT asset management companies in the US with 50-200 employees' --limit 100",
+        "autotouch search companies --query 'companies like Rippling in HR software, based in the United States, with 50-500 employees' --limit 100",
+        "autotouch search people --query 'VP IT' --company 'Rippling' --limit 10",
         "autotouch search recipe --type companies --out-file companies.json",
     ],
 }
@@ -141,11 +144,12 @@ _SEARCH_CAPABILITIES: Dict[str, Any] = {
 
 _SEARCH_RECIPES: Dict[str, Dict[str, Any]] = {
     "companies": {
-        "payload": {"query": "remote IT asset management", "limit": 10},
+        "payload": {"query": "remote IT asset management companies in the US with 50-200 employees", "limit": 100},
         "usage": "autotouch search companies --data-file <payload.json>",
         "notes": [
-            "Use this when you need to discover target accounts first.",
-            "Pricing: 1 credit per neural search API call. Max 10 returned companies per call.",
+            "Use this for basic Exa Company Search: company discovery, competitors, lookalikes, and semantic market scans.",
+            "Put Exa company constraints in natural language: industry, geography, funding, headcount, technology, founded date, and similarity.",
+            "Pricing: 1 credit per 10 returned Exa company results. Max 100 returned companies per call.",
         ],
     },
     "people": {
@@ -164,16 +168,15 @@ _SEARCH_RECIPES: Dict[str, Dict[str, Any]] = {
     },
     "similar_companies": {
         "payload": {
-            "company_name": "Rippling",
-            "company_domain": "rippling.com",
-            "company_description": "Workforce management and HR software platform for businesses",
-            "limit": 10,
+            "query": "companies like Rippling for workforce management, HR, payroll, IT, benefits, devices, and app access",
+            "limit": 100,
         },
         "usage": "autotouch search similar-companies --data-file <payload.json>",
         "notes": [
-            "Good for expanding a seed account list.",
-            "Best results come from providing company_domain and a short company_description when known.",
-            "Pricing: 1 credit per neural search API call. Max 10 returned companies per call.",
+            "Convenience wrapper for Exa Company Search similarity queries, not Exa /findSimilar.",
+            "Prefer query for Exa's documented semantic patterns: companies like X, competitors of X, or companies working on Y.",
+            "Seed fields company_name, company_domain, and company_description remain available when you want the CLI/API to compose the query.",
+            "Pricing: 1 credit per 10 returned Exa company results. Max 100 returned companies per call.",
         ],
     },
     "news": {

{autotouch_cli-0.2.85 → autotouch_cli-0.2.87}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "autotouch-cli"
-version = "0.2.85"
+version = "0.2.87"
 description = "Autotouch Smart Table CLI"
 readme = "README.md"
 requires-python = ">=3.9"