scrapeloop-mcp 0.7.1 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -2,10 +2,10 @@
2
2
 
3
3
  A [Model Context Protocol](https://modelcontextprotocol.io) server for Scrapeloop.
4
4
  **Set up and run Scrapeloop end-to-end from Claude Desktop, Cursor, ChatGPT, or any
5
- MCP client**connect integrations, scrape or pull from the lead database, verify
5
+ MCP client** - connect integrations, scrape or pull from the lead database, verify
6
6
  (incl. catch-all), clean, enrich, tag, and feed Instantly campaigns.
7
7
 
8
- Pair it with the [`scrapeloop-setup` skill](../skill) for a guided, step-by-step
8
+ Pair it with the [`scrapeloop` skill](../skill) for a saved Playbook setup and task
9
9
  flow ("set up Scrapeloop").
10
10
 
11
11
  ## Setup
@@ -52,7 +52,7 @@ integrations:write scrape enrich cleaners strategies campaigns:write insta
52
52
  ```
53
53
 
54
54
  A 403 with `"API key missing scope(s): …"` means the key was minted with a narrower
55
- setregenerate it with the full preset.
55
+ set - regenerate it with the full preset.
56
56
 
57
57
  ## Tools
58
58
 
@@ -64,7 +64,7 @@ cleaners, strategy tags, vendors, scopes, and the ordered 7 setup steps) and
64
64
  |------|-------|
65
65
  | Discovery | `get_capabilities`, `get_setup_status`, `get_credits`, `get_meta` |
66
66
  | Integrations (BYOK) | `list_integrations`, `create_integration`, `list_credentials`, `add_credential`, `test_credential`, `list_instantly_connections` |
67
- | Scrape | `list_scrapers`, `estimate_scrape`, `submit_scrape`, `list_jobs`, `get_job`, `cancel_job` |
67
+ | Scrape | `list_scrapers`, `estimate_scrape`, `submit_scrape`, `list_jobs`, `get_job`, `cancel_job`, `attach_job_to_table` |
68
68
  | Lead database | `search_leads`, `estimate_leads`, `reveal_lead`, `save_leads`, `bulk_verify_leads` |
69
69
  | Verify | `verify_email`, `verify_catchall` |
70
70
  | Clean | `list_cleaners`, `preview_cleaners`, `apply_cleaners` |
package/package.json CHANGED
@@ -1,10 +1,10 @@
1
1
  {
2
2
  "name": "scrapeloop-mcp",
3
- "version": "0.7.1",
4
- "description": "Scrapeloop MCP serverset up and run Scrapeloop end-to-end (integrations, scraping, lead database, verification, cleaners, enrichment, strategies, Instantly campaigns) from any MCP client (Claude Desktop, Cursor, ChatGPT, ).",
3
+ "version": "0.8.0",
4
+ "description": "Scrapeloop MCP server - set up and run Scrapeloop end-to-end (integrations, scraping, lead database, verification, cleaners, enrichment, strategies, Instantly campaigns) from any MCP client (Claude Desktop, Cursor, ChatGPT, \u2026).",
5
5
  "type": "module",
6
6
  "scripts": {
7
- "test": "node --check src/index.js && node test/mcp_table_tools.mjs && node test/mcp_community_intent.mjs && node test/mcp_api_key_preflight.mjs && node test/mutation_retry_contract.mjs && node test/release_readiness.mjs",
7
+ "test": "node --check src/index.js && node test/mcp_table_tools.mjs && node test/mcp_playbook_plans.mjs && node test/mcp_community_intent.mjs && node test/mcp_api_key_preflight.mjs && node test/mutation_retry_contract.mjs && node test/release_readiness.mjs",
8
8
  "release:preflight": "node scripts/release-preflight.mjs",
9
9
  "release:verify": "node scripts/release-installed-verify.mjs"
10
10
  },
@@ -21,10 +21,26 @@
21
21
  "dependencies": {
22
22
  "@modelcontextprotocol/sdk": "^1.0.0"
23
23
  },
24
- "keywords": ["mcp", "scrapeloop", "leads", "b2b", "email-verification", "cold-email", "model-context-protocol"],
24
+ "keywords": [
25
+ "mcp",
26
+ "scrapeloop",
27
+ "leads",
28
+ "b2b",
29
+ "email-verification",
30
+ "cold-email",
31
+ "model-context-protocol"
32
+ ],
25
33
  "homepage": "https://scrapeloop.com",
26
- "repository": { "type": "git", "url": "git+https://github.com/Hinkam-Lanello/scrapeloop.git", "directory": "packages/mcp" },
27
- "bugs": { "url": "https://github.com/Hinkam-Lanello/scrapeloop/issues" },
28
- "publishConfig": { "access": "public" },
34
+ "repository": {
35
+ "type": "git",
36
+ "url": "git+https://github.com/Hinkam-Lanello/scrapeloop.git",
37
+ "directory": "packages/mcp"
38
+ },
39
+ "bugs": {
40
+ "url": "https://github.com/Hinkam-Lanello/scrapeloop/issues"
41
+ },
42
+ "publishConfig": {
43
+ "access": "public"
44
+ },
29
45
  "license": "UNLICENSED"
30
46
  }
package/src/index.js CHANGED
@@ -1,19 +1,18 @@
1
1
  #!/usr/bin/env node
2
2
  /**
3
- * scrapeloop-mcp Model Context Protocol server for Scrapeloop.
3
+ * scrapeloop-mcp: Model Context Protocol server for Scrapeloop.
4
4
  *
5
5
  * A thin wrapper over the Scrapeloop REST API (/api/v1). Authenticates with a
6
6
  * per-workspace API key (Settings → API access). Credit-gated actions surface
7
7
  * 402/403 bodies verbatim so the model can self-correct or prompt a top-up.
8
8
  *
9
- * The full setup flow (connect choose source → configure → verify → clean →
10
- * enrich campaigns) is driven by these tools. Call `get_capabilities` first to
11
- * learn what this workspace actually has and `get_setup_status` to know where the
12
- * user already is, then walk the manifest's `steps` in order.
9
+ * Read get_capabilities, get_setup_status and get_playbook first. Use this
10
+ * workspace's saved answers for setup and later tasks, following the live
11
+ * manifest steps and each tool's approval and retry rules.
13
12
  *
14
13
  * Env:
15
- * SCRAPELOOP_API_KEY required sl_live_… key
16
- * SCRAPELOOP_API_URL optional defaults to https://api.scrapeloop.com
14
+ * SCRAPELOOP_API_KEY required: sl_live_… key
15
+ * SCRAPELOOP_API_URL optional, defaults to https://api.scrapeloop.com
17
16
  */
18
17
  import { Server } from '@modelcontextprotocol/sdk/server/index.js';
19
18
  import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
@@ -31,6 +30,21 @@ const TIMEOUT_MS = Number(process.env.SCRAPELOOP_TIMEOUT_MS) || 30000;
31
30
  const MAX_RETRIES = 3;
32
31
  const MAX_TABLE_CELL_WRITES = 500;
33
32
  const MAX_TABLE_CELL_REQUEST_BYTES = 256 * 1024;
33
+ // Native Smart Table tool keys, mirroring apps/api/lib/native_tool_columns.json.
34
+ // This package ships standalone, so it cannot read that file; the pairing is
35
+ // asserted in apps/web/lib/native-tool-columns.test.ts. Drift fails SAFE: the
36
+ // API rejects an unknown key and returns available_tools, and
37
+ // get_table_native_tools always serves the live catalog.
38
+ // verify_catchall is deliberately absent: its correctness needs a sibling Verify
39
+ // email column that has already RUN and returned Catch All or Unknown, which one
40
+ // create call cannot establish. Customers use verify_catchall (per address) or
41
+ // the in-app add-column menu. See docs/roadmap/mcp-native-tool-columns.md.
42
+ const NATIVE_TABLE_TOOLS = Object.freeze([
43
+ 'verify_email',
44
+ 'verify_phone',
45
+ 'email_finder',
46
+ 'business_email_finder',
47
+ ]);
34
48
  const TRACE_HEADER = 'X-Scrapeloop-Trace-Id';
35
49
  const SAFE_TRACE_ID = /^[A-Za-z0-9_-]{8,80}$/;
36
50
  const STARTUP_WARNINGS = new Set();
@@ -338,13 +352,34 @@ const runScopeBody = (a) => ({
338
352
  ...(a.only_failed !== undefined ? { only_failed: !!a.only_failed } : {}),
339
353
  });
340
354
 
355
+ // Plan choices mirror the app wizard. Explicit scope and limits prevent a
356
+ // missing answer from silently becoming a country-wide paid plan.
357
+ const PLAN_FIELDS = {
358
+ name: S, kind: S, integration_id: S, country: S,
359
+ granularity: { ...S, enum: ['country', 'state', 'city', 'postal_code'] },
360
+ locations: ARR(S), location_items: ARR(O), categories: ARR(S),
361
+ min_population: { type: 'integer', minimum: 0 }, scrape_config: O,
362
+ run_mode: { ...S, enum: ['all_now', 'over_time'] }, spread_evenly: B,
363
+ monthly_max_leads: { type: 'integer', minimum: 1 },
364
+ monthly_lead_ceiling: { type: 'integer', minimum: 1 },
365
+ monthly_max_usd: { ...N, exclusiveMinimum: 0 },
366
+ monthly_max_credits: { type: 'integer', minimum: 1 },
367
+ spend_buffer_usd: { ...N, minimum: 0 },
368
+ max_leads_per_chunk: { type: 'integer', minimum: 0 },
369
+ chunk_interval_minutes: { type: 'integer', minimum: 0 },
370
+ chunk_order: { ...S, enum: ['sequential', 'largest_first', 'balanced'] },
371
+ rescrape_covered: B, min_age_days: { type: 'integer', minimum: 1 },
372
+ rescrape_after_days: { type: 'integer', minimum: 1 },
373
+ };
374
+ const PLAN_REQUIRED = ['name', 'kind', 'country', 'granularity', 'locations', 'categories', 'run_mode'];
375
+
341
376
  // --- Tool registry: name → { def, run } -------------------------------------
342
377
  const TOOLS = {
343
378
  // ── Discovery / status ────────────────────────────────────────────────
344
379
  get_capabilities: {
345
380
  def: {
346
381
  description:
347
- 'Fetch the live Scrapeloop capability manifest: scrapers (+ their config schemas), enrichment primitives & presets, cleaners, strategy tags, vendors, scopes, filter enums, and the ordered 7 setup steps with guidance. CALL THIS FIRST and whenever unsure what is available it reflects the workspace as it is right now.',
382
+ 'Fetch the live Scrapeloop capability manifest: scrapers (+ their config schemas), enrichment primitives & presets, cleaners, strategy tags, vendors, scopes, filter enums, and the ordered Playbook setup and run steps with guidance. CALL THIS FIRST and whenever unsure what is available: it reflects the workspace as it is right now.',
348
383
  inputSchema: obj({}),
349
384
  },
350
385
  run: () => api('GET', '/manifest'),
@@ -352,7 +387,7 @@ const TOOLS = {
352
387
  get_setup_status: {
353
388
  def: {
354
389
  description:
355
- 'Onboarding checklist booleans (integration connected, a lead revealed, a verify run, a list created) so you know what is already done and where to resume.',
390
+ 'Onboarding checklist booleans (Playbook saved, integration connected, a lead revealed, a verify run, a Table created) so you know what is already done and where to resume.',
356
391
  inputSchema: obj({}),
357
392
  },
358
393
  run: () => api('GET', '/setup/status'),
@@ -402,7 +437,7 @@ const TOOLS = {
402
437
  },
403
438
  list_credentials: {
404
439
  def: {
405
- description: "List an integration's stored credentials (masked keys are never returned).",
440
+ description: "List an integration's stored credentials (masked: keys are never returned).",
406
441
  inputSchema: obj({ integration_id: S }, ['integration_id']),
407
442
  },
408
443
  run: (a) =>
@@ -442,7 +477,7 @@ const TOOLS = {
442
477
  list_scrapers: {
443
478
  def: {
444
479
  description:
445
- 'List available scrapers, whether each requires a credential, and the JSON Schema for each config. Ask the user the schema\'s fields (query, state, locations, limit, …), then fill `config` for estimate_scrape / submit_scrape.',
480
+ 'List available scrapers, whether each requires a credential, and the JSON Schema for each config. Reuse saved Playbook choices and ask only for missing fields. Read each plan capability before choosing preview_scrape_plan or a one-off estimate_scrape.',
446
481
  inputSchema: obj({}),
447
482
  },
448
483
  run: () => api('GET', '/scrapers'),
@@ -470,7 +505,7 @@ const TOOLS = {
470
505
  submit_scrape: {
471
506
  def: {
472
507
  description:
473
- 'Submit a scrape job. Optionally pass table_id from get_tables to add results to an existing writable static Table; omit it to keep results in All leads only. Paid scrapers spend vendor credits (BYOK) or lead credits (managed: 1 credit per NEW lead, deduped free), so confirm paid work with the user first. For every paid scraper, pass a positive hard_max_cost_usd equal to or above the estimate after the user confirms that ceiling; paid work will not start without it. A scraper with requires_credential=false is free and needs no integration_id, credential_id, or hard maximum. billing: "auto" (default) uses the workspace key when one exists, else Scrapeloop\'s managed key; "byok"/"managed" force the mode. integration_id is required for normal BYOK, optional for managed, and omitted for credential-free sources. Surfaces 402 (budget/free-tier/credits) and 409 (rescrape confirmation needed) verbatim; pass confirm_rescrape:true to proceed past a coverage conflict.',
508
+ 'Submit a scrape job. Optionally pass table_id from get_tables to add results to an existing writable static Table; omit it to keep results in All leads only. Paid scrapers spend vendor credits (BYOK) or lead credits (managed: 1 credit per NEW lead, deduped free), so confirm paid work with the user first. For every paid scraper, pass a positive hard_max_cost_usd equal to or above the estimate after the user confirms that ceiling; paid work will not start without it. A scraper with requires_credential=false is free and needs no integration_id, credential_id, or hard maximum. billing: "auto" (default) uses the workspace key when one exists, else Scrapeloop\'s managed key; "byok"/"managed" force the mode. integration_id is required for normal BYOK, optional for managed, and omitted for credential-free sources. Returns 402 for budget or credit limits and 409 when rescrape confirmation is needed; pass confirm_rescrape:true to proceed past a coverage conflict.',
474
509
  inputSchema: obj(
475
510
  {
476
511
  kind: S,
@@ -497,18 +532,55 @@ const TOOLS = {
497
532
  ...(a.billing ? { billing: a.billing } : {}),
498
533
  }),
499
534
  },
535
+ get_scrape_plan_territories: {
536
+ def: {
537
+ description: 'Read places a scraper can cover before previewing a plan. Use returned values exactly. For Outscraper, pass a returned parent value to read its child cities.',
538
+ inputSchema: obj({ kind: S, country: S, granularity: S, min_population: N, parent: S }, ['kind', 'country']),
539
+ },
540
+ run: (a) => api('GET', `/plans/territories?${new URLSearchParams(Object.entries(a).filter(([, value]) => value !== undefined)).toString()}`),
541
+ },
542
+ preview_scrape_plan: {
543
+ def: {
544
+ description: 'Preview all-now or over-time scraping with the same limits as the app. No scrape or credential is created. Read the Playbook and scraper capabilities first. Set a lead or credit cap for credit-priced work, or a dollar cap for dollar-priced work. all_now limits apply once; over_time limits reset each month. Show costs and coverage warnings before creating.',
545
+ inputSchema: obj(PLAN_FIELDS, PLAN_REQUIRED),
546
+ },
547
+ run: (a) => api('POST', '/plans/preview', a),
548
+ },
549
+ create_scrape_plan: {
550
+ def: {
551
+ description: 'Create and start an approved scrape plan into an existing writable Table. This can spend credits or vendor money. Preview the exact choices first and stay within the approved limit and time period. Pass explicit territory values and matching billing limits. Reuse the same idempotency_key after a lost reply, inspect list_scrape_plans and never create a second plan to escape an uncertain result. Over-time limits reset monthly; a total one-time approval does not authorize recurring spend.',
552
+ inputSchema: obj({ ...PLAN_FIELDS, target_list_id: { ...S, format: 'uuid' } }, [...PLAN_REQUIRED, 'target_list_id']),
553
+ },
554
+ run: ({ idempotency_key: _key, ...body }) => api('POST', '/plans', body),
555
+ },
556
+ list_scrape_plans: {
557
+ def: { description: 'Read this workspace scrape plans, progress and spending. Use after an uncertain plan create.', inputSchema: obj({}) },
558
+ run: () => api('GET', '/plans'),
559
+ },
560
+ get_scrape_plan: {
561
+ def: { description: 'Read one scrape plan, its saved scope, limits, progress and spending.', inputSchema: obj({ plan_id: { ...S, format: 'uuid' } }, ['plan_id']) },
562
+ run: (a) => api('GET', `/plans/${enc(a.plan_id)}`),
563
+ },
564
+ pause_scrape_plan: {
565
+ def: { description: 'Pause new work from a scrape plan. A job already running may still finish. Inspect the plan and its jobs after pausing.', inputSchema: obj({ plan_id: { ...S, format: 'uuid' } }, ['plan_id']) },
566
+ run: (a) => api('POST', `/plans/${enc(a.plan_id)}/pause`),
567
+ },
500
568
  list_jobs: {
501
569
  def: { description: 'List recent scrape/enrich jobs in the workspace.', inputSchema: obj({}) },
502
570
  run: () => api('GET', '/jobs'),
503
571
  },
504
572
  get_job: {
505
- def: { description: 'Get one job with its execution steps (poll this for scrape progress).', inputSchema: obj({ job_id: S }, ['job_id']) },
573
+ def: { description: 'Get one job with its execution steps (poll this for scrape progress). A scrape job carries `delivery`: how many rows were found, kept, new, already yours, and added to its Table, why rows were left out, and whether adding them to the Table worked.', inputSchema: obj({ job_id: S }, ['job_id']) },
506
574
  run: (a) => api('GET', `/jobs/${enc(a.job_id)}`),
507
575
  },
508
576
  cancel_job: {
509
577
  def: { description: 'Request cancellation of a queued or running job. An active external vendor task remains nonterminal while the worker aborts it and settles final partial usage, then becomes cancelled.', inputSchema: obj({ job_id: S }, ['job_id']) },
510
578
  run: (a) => api('POST', `/jobs/${enc(a.job_id)}/cancel`),
511
579
  },
580
+ attach_job_to_table: {
581
+ def: { description: 'Add a finished scrape job\'s leads to its Table again. Use it only when get_job shows `delivery.attach.state` as "failed" and the job is done, and not when `delivery.attach.code` is "table_missing" (that Table was deleted). Rows already in the Table are skipped, so repeating it never adds a row twice. Returns `added`, `in_table`, and the updated `delivery`.', inputSchema: obj({ job_id: S }, ['job_id']) },
582
+ run: (a) => api('POST', `/jobs/${enc(a.job_id)}/attach-to-table`),
583
+ },
512
584
 
513
585
  // ── Lead database ─────────────────────────────────────────────────────
514
586
  search_leads: {
@@ -533,7 +605,7 @@ const TOOLS = {
533
605
  },
534
606
  reveal_lead: {
535
607
  def: {
536
- description: "Reveal a lead's email by global_person_id. SPENDS 1 lead credit (idempotent already-revealed leads are free).",
608
+ description: "Reveal a lead's email by global_person_id. SPENDS 1 lead credit (idempotent: already-revealed leads are free).",
537
609
  inputSchema: obj({ global_person_id: S }, ['global_person_id']),
538
610
  },
539
611
  run: (a) => api('POST', '/leads/reveal', { global_person_id: a.global_person_id }),
@@ -678,7 +750,7 @@ const TOOLS = {
678
750
  verify_catchall: {
679
751
  def: {
680
752
  description:
681
- 'Resolve a mailbox on a catch-all domain with Scrapeloop managed verification (5 verify credits). High ROI: deliverable catch-all/unknown contacts get fewer cold emails because most senders skip them, so they reply more.',
753
+ 'Resolve a mailbox on a catch-all domain with Scrapeloop managed verification (5 verify credits). Optional and off unless chosen. It can resolve some unknown addresses but does not promise more replies. Estimate the eligible set and stay within the approved credit budget.',
682
754
  inputSchema: obj({ email: S }, ['email']),
683
755
  },
684
756
  run: (a) => api('POST', '/verify/catchall', { email: a.email }),
@@ -737,7 +809,7 @@ const TOOLS = {
737
809
  },
738
810
  run_enrich: {
739
811
  def: {
740
- description: 'Enqueue a bulk enrichment run. SPENDS vendor credits estimate + confirm first.',
812
+ description: 'Enqueue a bulk enrichment run. SPENDS vendor credits: estimate + confirm first.',
741
813
  inputSchema: obj({ preset_slug: S, preset_id: S, lead_ids: ARR(S), filter_query: O }),
742
814
  },
743
815
  run: (a) =>
@@ -768,6 +840,24 @@ const TOOLS = {
768
840
  ...(a.cache_settings ? { cache_settings: a.cache_settings } : {}),
769
841
  }),
770
842
  },
843
+ get_playbook: {
844
+ def: { description: "Read this workspace's saved Playbook and niches before setting up work.", inputSchema: obj({}) },
845
+ run: () => api('GET', '/playbook'),
846
+ },
847
+ update_playbook: {
848
+ def: {
849
+ description: 'Save defaults for this workspace only. Saves the sections you provide and keeps other answers. Never starts a scrape or spends credits.',
850
+ inputSchema: obj({ sections: { type: 'object', properties: Object.fromEntries(['channels', 'scrape', 'tables', 'checks', 'routing', 'campaigns'].map((key) => [key, { type: 'object' }])), additionalProperties: false } }, ['sections']),
851
+ },
852
+ run: (a) => api('PUT', '/playbook', a.sections),
853
+ },
854
+ set_playbook_niches: {
855
+ def: {
856
+ description: 'Save chosen niches in tiers 1, 2 or 3. A null tier removes a niche without deleting its category. Set replace to true only to replace all chosen niches.',
857
+ inputSchema: obj({ niches: ARR(obj({ gcid: { type: ['string', 'null'] }, name: { type: ['string', 'null'] }, tier: { type: ['integer', 'null'], enum: [1, 2, 3, null] } })), replace: B }, ['niches']),
858
+ },
859
+ run: (a) => api('PUT', '/playbook/niches', a),
860
+ },
771
861
  get_ai_context: {
772
862
  def: {
773
863
  description: 'Get the workspace AI context (company description, ICP, buyer personas) that seeds every AI feature.',
@@ -822,7 +912,7 @@ const TOOLS = {
822
912
  list_senders: {
823
913
  def: {
824
914
  description:
825
- 'List the sender vendors Scrapeloop supports (instantly, …) with each one\'s capabilities (webhooks, lead lists, async moves, bulk upload). Catalog read use it to know what a connected sender can do before configuring feed/offload.',
915
+ 'List the sender vendors Scrapeloop supports (instantly, …) with each one\'s capabilities (webhooks, lead lists, async moves, bulk upload). Catalog read: use it to know what a connected sender can do before configuring feed/offload.',
826
916
  inputSchema: obj({}),
827
917
  },
828
918
  run: () => api('GET', '/senders'),
@@ -830,7 +920,7 @@ const TOOLS = {
830
920
  list_sender_campaigns: {
831
921
  def: {
832
922
  description:
833
- "List a sender credential's campaigns at the vendor (vendor-generic: works for any connected sender key, same shape as list_instantly_campaigns use it to pick external_list_id).",
923
+ "List a sender credential's campaigns at the vendor (vendor-generic: works for any connected sender key, same shape as list_instantly_campaigns; use it to pick external_list_id).",
834
924
  inputSchema: obj({ credential_id: S }, ['credential_id']),
835
925
  },
836
926
  run: (a) => api('GET', `/senders/${enc(a.credential_id)}/campaigns`),
@@ -867,6 +957,10 @@ const TOOLS = {
867
957
  def: { description: 'List campaigns with per-campaign stats and the bound list name.', inputSchema: obj({}) },
868
958
  run: () => api('GET', '/campaigns'),
869
959
  },
960
+ get_campaign_setup_options: {
961
+ def: { description: 'Read this workspace\'s channel choices and defaults for new campaign drafts. Explicit create choices win. Reading options changes nothing and starts no campaign.', inputSchema: obj({}) },
962
+ run: () => api('GET', '/campaigns/options'),
963
+ },
870
964
  preview_add_leads_to_campaign: {
871
965
  def: {
872
966
  description:
@@ -894,21 +988,32 @@ const TOOLS = {
894
988
  create_campaign: {
895
989
  def: {
896
990
  description:
897
- 'Create a campaign binding a Scrapeloop list (source_list_id) OR segment to an Instantly campaign (external_list_id from list_instantly_campaigns). feed_mode: off|auto_add|drip. Set target_active_count, cooldown_days (default 180), resting_period_days (default 90). Created paused confirm before activate_campaign.',
991
+ 'Create a draft campaign connected to a sender and a Scrapeloop Table or segment. Read get_campaign_setup_options first. Absent choices use this workspace Playbook; explicit values win. GoHighLevel SMS needs a source Table and external_list_id set to a workflow ID or contacts-only. Nothing starts until activate_campaign. Ask the user before Start.',
898
992
  inputSchema: obj(
899
993
  {
900
994
  name: S,
995
+ sender_options: { ...O, description: "GoHighLevel SMS choices: workflow_id, tags, require_verified_mobile, voip_counts_as_mobile, missing_fields (ask, create, never)." },
996
+ icon: { type: ["string", "null"], description: "One or two emoji, or null to clear." },
901
997
  source_list_id: S,
902
998
  segment_id: S,
903
999
  sender_credential_id: S,
904
1000
  external_list_id: S,
905
1001
  instantly_campaign_name: S,
906
1002
  feed_mode: { ...S, enum: ['off', 'auto_add', 'drip'] },
1003
+ allow_risky_emails: B,
1004
+ offload_enabled: B,
1005
+ offload_no_reply_days: N,
1006
+ offload_to_instantly: B,
1007
+ offload_to_scrapeloop: B,
1008
+ offload_scrapeloop_list_id: S,
907
1009
  drip_strategy: { ...S, enum: ['target_active', 'fixed_daily'] },
908
1010
  drip_daily_count: N,
909
1011
  target_active_count: N,
910
1012
  cooldown_days: N,
911
1013
  resting_period_days: N,
1014
+ recycle_mode: { ...S, enum: ['off', 'after_cooldown'] },
1015
+ prioritize_fresh: B,
1016
+ bounce_instantly_list_id: S,
912
1017
  },
913
1018
  ['name', 'sender_credential_id', 'external_list_id'],
914
1019
  ),
@@ -917,7 +1022,10 @@ const TOOLS = {
917
1022
  const body = { name: a.name, sender_credential_id: a.sender_credential_id, external_list_id: a.external_list_id };
918
1023
  for (const k of [
919
1024
  'source_list_id', 'segment_id', 'instantly_campaign_name', 'feed_mode', 'drip_strategy',
1025
+ 'allow_risky_emails', 'offload_enabled', 'offload_no_reply_days', 'offload_to_instantly',
1026
+ 'offload_to_scrapeloop', 'offload_scrapeloop_list_id',
920
1027
  'drip_daily_count', 'target_active_count', 'cooldown_days', 'resting_period_days',
1028
+ 'recycle_mode', 'prioritize_fresh', 'bounce_instantly_list_id', 'sender_options', 'icon',
921
1029
  ]) {
922
1030
  if (a[k] !== undefined && a[k] !== null) body[k] = a[k];
923
1031
  }
@@ -925,19 +1033,19 @@ const TOOLS = {
925
1033
  },
926
1034
  },
927
1035
  update_campaign: {
928
- def: { description: 'Patch a campaign (feed_mode, target_active_count, cooldown_days, external_list_id, …).', inputSchema: obj({ campaign_id: S, patch: O }, ['campaign_id', 'patch']) },
1036
+ def: { description: 'Patch a campaign (feed_mode, target_active_count, cooldown_days, external_list_id, …). Switching feed_mode on for an already active campaign runs the same check as campaign_readiness_check: any fail answers 422 with the report under detail.readiness.', inputSchema: obj({ campaign_id: S, patch: { ...O, description: "Campaign settings to change.", properties: { sender_options: { ...O, description: "GoHighLevel SMS settings. Only supplied keys change." }, icon: { type: ["string", "null"], description: "One or two emoji, or null to clear." } }, additionalProperties: true } }, ['campaign_id', 'patch']) },
929
1037
  run: (a) => api('PATCH', `/campaigns/${enc(a.campaign_id)}`, a.patch || {}),
930
1038
  },
931
1039
  duplicate_campaign: {
932
1040
  def: {
933
1041
  description:
934
- 'Duplicate a campaign: clones ALL configuration (list/segment binding, sender credential, Instantly binding, feed/drip/offload settings, field mapping) into a new draft named "Copy of …". Created SAFE regardless of the source feed_mode is forced to off and require_approval to true, so the copy never feeds leads to Instantly until explicitly enabled. Leads/stats/ledger are NOT copied.',
1042
+ 'Duplicate a campaign: clones ALL configuration (list/segment binding, sender credential, Instantly binding, feed/drip/offload settings, field mapping) into a new draft named "Copy of …". Created SAFE regardless of the source: feed_mode is forced to off and require_approval to true, so the copy never feeds leads to Instantly until explicitly enabled. Leads/stats/ledger are NOT copied.',
935
1043
  inputSchema: obj({ campaign_id: S }, ['campaign_id']),
936
1044
  },
937
1045
  run: (a) => api('POST', `/campaigns/${enc(a.campaign_id)}/duplicate`),
938
1046
  },
939
1047
  activate_campaign: {
940
- def: { description: 'Activate a campaign STARTS the feed/send loop. Confirm with the user first.', inputSchema: obj({ campaign_id: S }, ['campaign_id']) },
1048
+ def: { description: 'Activate a campaign: STARTS the feed/send loop. Run campaign_readiness_check first; any fail makes this answer 422 with the report under detail.readiness. Confirm with the user first.', inputSchema: obj({ campaign_id: S }, ['campaign_id']) },
941
1049
  run: (a) => api('POST', `/campaigns/${enc(a.campaign_id)}/activate`),
942
1050
  },
943
1051
  pause_campaign: {
@@ -952,6 +1060,53 @@ const TOOLS = {
952
1060
  def: { description: "A campaign's ledger stats + best-effort live Instantly analytics.", inputSchema: obj({ campaign_id: S }, ['campaign_id']) },
953
1061
  run: (a) => api('GET', `/campaigns/${enc(a.campaign_id)}/stats`),
954
1062
  },
1063
+ get_campaign_field_mapping: {
1064
+ def: {
1065
+ description:
1066
+ 'What a campaign sends to the sender and what each detail is called there. Returns every name the sequence uses, every source that could fill it (with how full each one is over the first 500 rows), and the obvious matches. Read this when a merge variable comes out blank in the emails.',
1067
+ inputSchema: obj({ campaign_id: S }, ['campaign_id']),
1068
+ },
1069
+ run: (a) => api('GET', `/campaigns/${enc(a.campaign_id)}/field-mapping`),
1070
+ },
1071
+ create_sender_fields: {
1072
+ def: {
1073
+ description: 'Create missing GoHighLevel SMS contact fields only after the user agrees. Existing fields are reused. With remember never, save the choice and create nothing. Returns requested names mapped to actual field keys; use those keys when saving the mapping.',
1074
+ inputSchema: obj({ campaign_id: S, names: ARR(S), remember: { type: ['string', 'null'], enum: ['create', 'never', null] } }, ['campaign_id', 'names']),
1075
+ },
1076
+ run: (a) => api('POST', `/campaigns/${enc(a.campaign_id)}/sender-fields`, { names: a.names, remember: a.remember ?? null }),
1077
+ },
1078
+ set_campaign_field_mapping: {
1079
+ def: {
1080
+ description:
1081
+ 'Set what a campaign sends. `mapping` is {name the sender receives: where it comes from}, using lead.<column>, send.<name>, col.<table column key> or custom.<key>. `identity` may override first_name, last_name and company only, never email. `only_mapped` true means the mapping is the complete list and nothing else is sent. Saving also records that a person reviewed it, which a campaign needs before its feed can be switched on.',
1082
+ inputSchema: obj(
1083
+ { campaign_id: S, mapping: { type: 'object' }, identity: { type: 'object' }, only_mapped: B },
1084
+ ['campaign_id'],
1085
+ ),
1086
+ },
1087
+ run: (a) =>
1088
+ api('PUT', `/campaigns/${enc(a.campaign_id)}/field-mapping`, {
1089
+ mapping: a.mapping || {},
1090
+ identity: a.identity || {},
1091
+ only_mapped: a.only_mapped !== false,
1092
+ }),
1093
+ },
1094
+ campaign_delivery_coverage: {
1095
+ def: {
1096
+ description:
1097
+ 'What the live sequence asks for versus what the feed will actually supply. Names any merge variable that nothing fills, which goes out blank in the middle of a sentence, and any that is only filled on some leads.',
1098
+ inputSchema: obj({ campaign_id: S }, ['campaign_id']),
1099
+ },
1100
+ run: (a) => api('GET', `/campaigns/${enc(a.campaign_id)}/delivery-coverage`),
1101
+ },
1102
+ campaign_readiness_check: {
1103
+ def: {
1104
+ description:
1105
+ 'Is this campaign ready to start? Run it before activate_campaign. Returns ready, sender_known, sampled and a list of checks, each with a stable key, a level (pass, warn, fail or info), one plain sentence and a fix. Any fail means activate_campaign will refuse with 422 and the same report under detail.readiness. A warn does not block: read it out to the user before they confirm. It judges the leads the feed sends next (not the ones already in the campaign). With feed_mode off nothing is uploaded, so the lead and mailbox checks only warn until the feed is switched on.',
1106
+ inputSchema: obj({ campaign_id: S }, ['campaign_id']),
1107
+ },
1108
+ run: (a) => api('GET', `/campaigns/${enc(a.campaign_id)}/readiness`),
1109
+ },
955
1110
  get_sender_config: {
956
1111
  def: { description: "Get an Instantly credential's sender config (daily cap, send accounts, …).", inputSchema: obj({ credential_id: S }, ['credential_id']) },
957
1112
  run: (a) => api('GET', `/sender_configs/${enc(a.credential_id)}`),
@@ -1004,6 +1159,35 @@ const TOOLS = {
1004
1159
  },
1005
1160
  run: () => api('GET', '/workbench/tree'),
1006
1161
  },
1162
+ get_pipeline: {
1163
+ def: {
1164
+ description: 'Read the saved result and progress for a pipeline. Use its operation_id after an interrupted create.',
1165
+ inputSchema: obj({ operation_id: { ...S, format: 'uuid' } }, ['operation_id']),
1166
+ },
1167
+ run: (a) => api('GET', `/pipelines/${enc(a.operation_id)}`),
1168
+ },
1169
+ preview_pipeline: {
1170
+ def: {
1171
+ description: 'Preview an empty pipeline from this workspace Playbook, including Tables, columns, checks, rules and future-row work. Creates nothing. Review this before create_pipeline.',
1172
+ inputSchema: obj({}),
1173
+ },
1174
+ run: (a) => api('POST', '/pipelines/preview', a),
1175
+ },
1176
+ create_pipeline: {
1177
+ def: {
1178
+ description: 'Create an empty pipeline from this workspace Playbook. Raw keeps originals. Rules start on, and chosen columns run on future rows. Creation starts no scrape, sends no messages and activates no campaign. Read get_playbook first. Reuse the same operation_id and choices to resume a partial create. Campaign bindings require campaigns:write and feeds off.',
1179
+ inputSchema: obj({
1180
+ operation_id: { ...S, format: 'uuid' }, name: S, market: S,
1181
+ niche_gcids: ARR(S), mode: { ...S, enum: ['auto', 'simple', 'advanced'] },
1182
+ channels: ARR({ ...S, enum: ['email', 'sms', 'linkedin'] }),
1183
+ checks: { ...O, description: 'Overrides for saved checks, using the Playbook checks keys.' },
1184
+ stage_mode: { ...S, enum: ['move', 'copy'] }, carry_values: B,
1185
+ raw_setup: { ...O, description: 'Optional Raw columns, packs and layout. Same shape as create_table table_setup.' },
1186
+ bindings: { ...O, description: 'Optional email, sms or linkedin campaign IDs. Existing campaign feeds must be off.' },
1187
+ }, ['operation_id', 'name']),
1188
+ },
1189
+ run: (a) => api('POST', '/pipelines', a),
1190
+ },
1007
1191
  create_workbook: {
1008
1192
  def: {
1009
1193
  description:
@@ -1016,7 +1200,7 @@ const TOOLS = {
1016
1200
  def: {
1017
1201
  description:
1018
1202
  'Add one tab to a workbook. Pass list_id to attach an existing standalone Table, or new_table_name to create a new empty Table and attach it. Pass exactly one.',
1019
- inputSchema: obj({ workbook_id: S, list_id: S, new_table_name: S }, ['workbook_id']),
1203
+ inputSchema: obj({ workbook_id: S, list_id: S, new_table_name: S, operation_id: { ...S, format: 'uuid' }, table_setup: { ...O, description: 'New Tables only. Same saved-column, pack and layout choices as create_table.' } }, ['workbook_id']),
1020
1204
  },
1021
1205
  run: (a) => {
1022
1206
  if (!!a.list_id === !!a.new_table_name) {
@@ -1029,6 +1213,8 @@ const TOOLS = {
1029
1213
  return api('POST', `/workbooks/${enc(a.workbook_id)}/tables`, {
1030
1214
  ...(a.list_id ? { list_id: a.list_id } : {}),
1031
1215
  ...(a.new_table_name ? { new_table_name: a.new_table_name } : {}),
1216
+ ...(a.operation_id ? { operation_id: a.operation_id } : {}),
1217
+ ...(a.table_setup !== undefined ? { table_setup: a.table_setup } : {}),
1032
1218
  });
1033
1219
  },
1034
1220
  },
@@ -1120,16 +1306,33 @@ const TOOLS = {
1120
1306
  },
1121
1307
  run: (a) => api('GET', `/tables${qs({ archived: a.archived })}`),
1122
1308
  },
1309
+ get_table_setup_options: {
1310
+ def: {
1311
+ description: 'Read this workspace\'s saved Table columns, layout and available column packs. Read-only; starts no work.',
1312
+ inputSchema: obj({}),
1313
+ },
1314
+ run: () => api('GET', '/playbook/tables/options'),
1315
+ },
1123
1316
  create_table: {
1124
1317
  def: {
1125
1318
  description:
1126
- 'Create a campaign-bindable Scrapeloop Table. Static is the default for imported/manual leads and accepts only name, description, and kind. filter_expression, sort_expression, visible_columns, and is_shared are smart-Table-only fields. The create is retry-safe: one operation_id is reused across HTTP retries. You may pass a prior operation_id to recover an ambiguous call; reusing it with different inputs returns 409.',
1319
+ 'Create a Scrapeloop Table. Regular Tables start with this workspace\'s saved columns and layout. Use table_setup to change columns, choose packs or rename labels. Creation never runs columns or spends credits. Smart Tables use filter_expression, sort_expression, visible_columns and is_shared, and cannot use table_setup. Retry with the same operation_id; different inputs return 409.',
1127
1320
  inputSchema: obj(
1128
1321
  {
1129
1322
  name: S,
1130
1323
  operation_id: { ...S, format: 'uuid', description: 'Optional retry key from a prior ambiguous call.' },
1131
1324
  description: S,
1132
1325
  kind: { ...S, enum: ['static', 'smart'] },
1326
+ table_setup: {
1327
+ ...O,
1328
+ description: 'Regular Tables only. Read get_table_setup_options first. columns replaces saved columns (empty list means none); packs adds named packs; identity_column_state replaces the saved layout. All created columns start with auto-run off.',
1329
+ properties: {
1330
+ columns: ARR({ ...O, properties: { kind: { ...S, enum: ['native', 'preset'] }, key: S, label: S, config: O }, required: ['kind', 'key', 'label'] }),
1331
+ packs: ARR(S),
1332
+ identity_column_state: O,
1333
+ },
1334
+ additionalProperties: false,
1335
+ },
1133
1336
  filter_expression: { ...O, description: 'Smart Tables only.' },
1134
1337
  sort_expression: { ...O, description: 'Smart Tables only.' },
1135
1338
  visible_columns: { ...ARR(S), description: 'Smart Tables only.' },
@@ -1157,6 +1360,7 @@ const TOOLS = {
1157
1360
  operation_id: a.operation_id || randomUUID(),
1158
1361
  ...(a.description !== undefined ? { description: a.description } : {}),
1159
1362
  ...(a.kind !== undefined ? { kind: a.kind } : {}),
1363
+ ...(a.table_setup !== undefined ? { table_setup: a.table_setup } : {}),
1160
1364
  ...(a.filter_expression !== undefined ? { filter_expression: a.filter_expression } : {}),
1161
1365
  ...(a.sort_expression !== undefined ? { sort_expression: a.sort_expression } : {}),
1162
1366
  ...(a.visible_columns !== undefined ? { visible_columns: a.visible_columns } : {}),
@@ -1338,6 +1542,86 @@ const TOOLS = {
1338
1542
  });
1339
1543
  },
1340
1544
  },
1545
+ get_table_native_tools: {
1546
+ def: {
1547
+ description:
1548
+ 'List the first-party Scrapeloop tools a Smart Table column can be built on, with their per-row credit price, documented options, and the legacy preset slugs each one supersedes. Read this when create_table_preset_column refuses a slug as superseded, or before create_table_native_column, so the tool key and options are exact rather than guessed.',
1549
+ inputSchema: obj({}, []),
1550
+ },
1551
+ run: () => api('GET', '/tables/native-tools'),
1552
+ },
1553
+ create_table_native_column: {
1554
+ def: {
1555
+ description:
1556
+ `Attach one native Scrapeloop tool as a real Smart Table column: ${NATIVE_TABLE_TOOLS.join(', ')}. These are the first-party tools that superseded four legacy presets, so they are addressed by tool key here rather than by preset slug through create_table_preset_column (which refuses a superseded slug for a new column). Each bills MANAGED VERIFICATION CREDITS per row, never dollars and never the customer's own vendor keys: verify_email reserves 1 credit per selected row, verify_phone reserves 5, and each email finder reserves 3. Settlement charges only fresh vendor checks and returns the unused reservation; missing inputs, cached rows, and skipped rows are free. Creating a column is always free. auto_run defaults false and queues nothing; when auto_run would backfill existing rows the API returns a 409 estimate whose cost.total_credits is the real charge, and you confirm it with confirm_auto_run plus confirmed_credits (CREDITS, not dollars - estimated_cost_usd is our vendor cost and is far smaller than what the workspace is billed). An auto_run that exceeds the workspace verify-credit balance returns 402 with {credit_kind, required, available} before anything is created. The Table master auto-run switch and workspace budget caps stay fail-closed. options are per tool and only the documented names are accepted - call get_table_native_tools for the current list. Owner email finder needs an owner first name plus the company website; point options.owner_name_column at an Owner name finder column to source the name, which auto_run requires. Business email finder is a SEPARATE column - create it with its own call and set options.owner_email_column to run it behind an owner finder. The Catch-all verifier is NOT creatable here: it needs a Verify email column that has already RUN and returned Catch All or Unknown, so use verify_catchall per address, or add the column from the Table's add-column menu.`,
1557
+ inputSchema: obj(
1558
+ {
1559
+ table_id: S,
1560
+ tool: { ...S, enum: NATIVE_TABLE_TOOLS },
1561
+ label: { ...S, description: 'Column label. Defaults to the tool name.' },
1562
+ auto_run: B,
1563
+ confirm_auto_run: B,
1564
+ confirmed_credits: {
1565
+ ...N,
1566
+ minimum: 0,
1567
+ description:
1568
+ 'The 409 estimate\'s cost.total_credits, echoed back exactly. Verification credits, not dollars.',
1569
+ },
1570
+ options: {
1571
+ ...O,
1572
+ description:
1573
+ 'Tool-specific options. verify_phone: phone_col, default_region. email_finder: owner_name_column. business_email_finder: owner_email_column. An undocumented name is rejected, never ignored.',
1574
+ },
1575
+ operation_id: {
1576
+ ...S,
1577
+ format: 'uuid',
1578
+ description: 'Required stable retry key for this logical native-column create.',
1579
+ },
1580
+ },
1581
+ ['table_id', 'tool', 'operation_id'],
1582
+ ),
1583
+ },
1584
+ run: async (a) => {
1585
+ if (!a.table_id || !a.tool || !a.operation_id) {
1586
+ return {
1587
+ ok: false,
1588
+ status: 400,
1589
+ error: { detail: 'table_id, tool, and a stable operation_id are required.' },
1590
+ };
1591
+ }
1592
+ if (!NATIVE_TABLE_TOOLS.includes(a.tool)) {
1593
+ return {
1594
+ ok: false,
1595
+ status: 400,
1596
+ error: {
1597
+ detail: `Unknown native tool '${a.tool}'. Expected one of ${NATIVE_TABLE_TOOLS.join(', ')}. Call get_table_native_tools for the live catalog.`,
1598
+ },
1599
+ };
1600
+ }
1601
+ const requestBody = {
1602
+ tool: a.tool,
1603
+ operation_id: a.operation_id,
1604
+ ...(a.label !== undefined ? { label: a.label } : {}),
1605
+ ...(a.auto_run !== undefined ? { auto_run: a.auto_run } : {}),
1606
+ ...(a.confirm_auto_run !== undefined ? { confirm_auto_run: a.confirm_auto_run } : {}),
1607
+ ...(a.confirmed_credits !== undefined
1608
+ ? { confirmed_credits: a.confirmed_credits }
1609
+ : {}),
1610
+ ...(a.options !== undefined ? { options: a.options } : {}),
1611
+ };
1612
+ return api('POST', `/tables/${enc(a.table_id)}/native-columns`, requestBody, {
1613
+ reconciliation: {
1614
+ operation_id: a.operation_id,
1615
+ retry_with: {
1616
+ tool: 'create_table_native_column',
1617
+ arguments: { ...a, operation_id: a.operation_id },
1618
+ },
1619
+ retry_guidance:
1620
+ `Retry create_table_native_column with the same operation_id ${a.operation_id}. Never substitute a new key for this logical create.`,
1621
+ },
1622
+ });
1623
+ },
1624
+ },
1341
1625
  set_table_cells: {
1342
1626
  def: {
1343
1627
  description:
@@ -1453,7 +1737,7 @@ const TOOLS = {
1453
1737
  preview_apify_import: {
1454
1738
  def: {
1455
1739
  description:
1456
- "Read-only preview of an approved Apify actor's already-produced dataset before importing it. mapping is {lead_field: dotted.source.path} where lead_field is one of name, email, phone, domain, organization_name, city, state, country, linkedin_url, title. Normalizes the whole dataset (rows with no email/phone/domain/linkedin/name are skipped), returns the first 25 rows plus record_count (the true total) and a preview_hash. Pass that preview_hash to import_apify_dataset the import re-verifies it and refuses (409) if the dataset changed. Does NOT run the actor or spend: it reads an existing dataset_id. Returns {dataset_id, actor_id, record_count, preview, preview_truncated, preview_hash}.",
1740
+ "Read-only preview of an approved Apify actor's already-produced dataset before importing it. mapping is {lead_field: dotted.source.path} where lead_field is one of name, email, phone, domain, organization_name, city, state, country, linkedin_url, title. Normalizes the whole dataset (rows with no email/phone/domain/linkedin/name are skipped), returns the first 25 rows plus record_count (the true total) and a preview_hash. Pass that preview_hash to import_apify_dataset: the import re-verifies it and refuses (409) if the dataset changed. Does NOT run the actor or spend: it reads an existing dataset_id. Returns {dataset_id, actor_id, record_count, preview, preview_truncated, preview_hash}.",
1457
1741
  inputSchema: obj(
1458
1742
  {
1459
1743
  dataset_id: { ...S, description: 'Apify dataset id from a completed actor run.' },
@@ -1468,7 +1752,7 @@ const TOOLS = {
1468
1752
  import_apify_dataset: {
1469
1753
  def: {
1470
1754
  description:
1471
- 'Import a previewed Apify dataset (≤1000 rows) into a private Scrapeloop Table through the bring-your-own-leads funnel (dedupe + identity matching + Table attach reused; no lead credits). Call preview_apify_import first and pass its preview_hash the import re-reads the dataset and returns 409 (preview_drift) if it changed since preview. Each row gets a stable external_id (apify:{dataset}:{row}) so re-importing updates rather than duplicates; company lands in business_name and the contact name/title ride through as custom fields. Target list_id, or omit for a new "Apify import" Table. Returns the import funnel result {inserted, updated, deduped, invalid, ...}.',
1755
+ 'Import a previewed Apify dataset (≤1000 rows) into a private Scrapeloop Table through the bring-your-own-leads funnel (dedupe + identity matching + Table attach reused; no lead credits). Call preview_apify_import first and pass its preview_hash: the import re-reads the dataset and returns 409 (preview_drift) if it changed since preview. Each row gets a stable external_id (apify:{dataset}:{row}) so re-importing updates rather than duplicates; company lands in business_name and the contact name/title ride through as custom fields. Target list_id, or omit for a new "Apify import" Table. Returns the import funnel result {inserted, updated, deduped, invalid, ...}.',
1472
1756
  inputSchema: obj(
1473
1757
  {
1474
1758
  dataset_id: S,
@@ -1646,7 +1930,7 @@ const TOOLS = {
1646
1930
  preview_import: {
1647
1931
  def: {
1648
1932
  description:
1649
- 'Dry-run for import_leads: report exactly what an import WOULD do {would_insert, would_update, deduped, invalid, invalid_reasons, list_exists} WITHOUT writing anything (no leads, no list, no tags created). Same inputs as import_leads. Call this first for a bring-your-own batch, show the user the plan (e.g. "adds 340 new, updates 12, skips 3 bad rows"), then confirm before import_leads.',
1933
+ 'Dry-run for import_leads: report exactly what an import WOULD do ({would_insert, would_update, deduped, invalid, invalid_reasons, list_exists}) WITHOUT writing anything (no leads, no list, no tags created). Same inputs as import_leads. Call this first for a bring-your-own batch, show the user the plan (e.g. "adds 340 new, updates 12, skips 3 bad rows"), then confirm before import_leads.',
1650
1934
  inputSchema: obj(
1651
1935
  {
1652
1936
  list_id: S,
@@ -2068,7 +2352,7 @@ const TOOLS = {
2068
2352
  list_table_rows: {
2069
2353
  def: {
2070
2354
  description:
2071
- "Read a table's rows the way the grid sees them lead identity fields (name/email/phone/domain/city/state/country/status) plus every column's cell (status + value). Server-side: sort is a JSON-array string like [{\"key\":\"lead.name\",\"dir\":\"asc\"},{\"key\":\"col.company_size\",\"dir\":\"desc\"}] (keys are lead.<field> or col.<column_key>, ≤3 levels); q is a full-text search across identity fields + visible column cells; filter is a FilterExpr object (same grammar the Tables filter UI uses) e.g. {\"and\":[{\"field\":\"lead.email_status\",\"op\":\"eq\",\"value\":\"valid\"},{\"field\":\"col.company_size\",\"op\":\"gte\",\"value\":50}]} and adds total_filtered alongside total; pass the previous response's next_cursor back as cursor to page (null next_cursor = last page). Distinct from get_list_rows (the plain import read-back). limit ≤ 500.",
2355
+ "Read a table's rows the way the grid sees them: lead identity fields (name/email/phone/domain/city/state/country/status) plus every column's cell (status + value). Server-side: sort is a JSON-array string like [{\"key\":\"lead.name\",\"dir\":\"asc\"},{\"key\":\"col.company_size\",\"dir\":\"desc\"}] (keys are lead.<field> or col.<column_key>, ≤3 levels); q is a full-text search across identity fields + visible column cells; filter is a FilterExpr object (same grammar the Tables filter UI uses), e.g. {\"and\":[{\"field\":\"lead.email_status\",\"op\":\"eq\",\"value\":\"valid\"},{\"field\":\"col.company_size\",\"op\":\"gte\",\"value\":50}]}, and adds total_filtered alongside total; pass the previous response's next_cursor back as cursor to page (null next_cursor = last page). Distinct from get_list_rows (the plain import read-back). limit ≤ 500.",
2072
2356
  inputSchema: obj(
2073
2357
  {
2074
2358
  list_id: S,
@@ -2076,7 +2360,7 @@ const TOOLS = {
2076
2360
  limit: N,
2077
2361
  sort: S,
2078
2362
  q: S,
2079
- filter: { ...O, description: 'FilterExpr JSON (object) server-side row filter; adds total_filtered.' },
2363
+ filter: { ...O, description: 'FilterExpr JSON (object): server-side row filter; adds total_filtered.' },
2080
2364
  view_id: { ...S, description: 'A saved custom view uuid or a system key (errored_rows/fully_enriched/data_only) from list_table_views. An ad-hoc filter/sort replaces the view\'s.' },
2081
2365
  },
2082
2366
  ['list_id'],
@@ -2100,7 +2384,7 @@ const TOOLS = {
2100
2384
  list_table_views: {
2101
2385
  def: {
2102
2386
  description:
2103
- "List a table's views three always-current SYSTEM views (errored_rows / fully_enriched / data_only) plus the user's saved custom views. Pass a returned id (or a system key) as view_id to list_table_rows to read that view (a saved filter + sorts + column overlay).",
2387
+ "List a table's views: three always-current SYSTEM views (errored_rows / fully_enriched / data_only) plus the user's saved custom views. Pass a returned id (or a system key) as view_id to list_table_rows to read that view (a saved filter + sorts + column overlay).",
2104
2388
  inputSchema: obj({ list_id: S }, ['list_id']),
2105
2389
  },
2106
2390
  run: (a) =>
@@ -2156,7 +2440,7 @@ const TOOLS = {
2156
2440
  update_table_view: {
2157
2441
  def: {
2158
2442
  description:
2159
- 'Update a saved table view any field (name/description/filter_expression/sorts/column_state/row_window); a field set to null clears it. System views (errored_rows/fully_enriched/data_only) are read-only.',
2443
+ 'Update a saved table view: any field (name/description/filter_expression/sorts/column_state/row_window); a field set to null clears it. System views (errored_rows/fully_enriched/data_only) are read-only.',
2160
2444
  inputSchema: obj(
2161
2445
  { list_id: S, view_id: S, name: S, description: S, filter_expression: O, sorts: ARR(O), column_state: O, row_window: O },
2162
2446
  ['list_id', 'view_id'],
@@ -2173,7 +2457,7 @@ const TOOLS = {
2173
2457
  delete_table_view: {
2174
2458
  def: {
2175
2459
  description:
2176
- 'Delete a saved table view rows in the table are unaffected. Confirm with the user first. System views cannot be deleted.',
2460
+ 'Delete a saved table view. Rows in the table are unaffected. Confirm with the user first. System views cannot be deleted.',
2177
2461
  inputSchema: obj({ list_id: S, view_id: S }, ['list_id', 'view_id']),
2178
2462
  },
2179
2463
  run: (a) =>
@@ -2184,7 +2468,7 @@ const TOOLS = {
2184
2468
  estimate_table_column: {
2185
2469
  def: {
2186
2470
  description:
2187
- "FREE cost preview for running one enrichment column, optionally scoped to a view or selection (view_id + selection{mode:'query', filter:<FilterExpr>, exclude_ids} or {mode:'explicit', ids}). cell_filter picks which cells: 'empty_or_stale' (default in the UI skips results still fresh under the current column config), 'failed', or 'all'; n_rows/start_row window the run (1-based). Returns lead_count + estimated_cost_usd + scope_resolved. ALWAYS show this to the user before run_table_column it spends nothing.",
2471
+ "FREE cost preview for running one enrichment column, optionally scoped to a view or selection (view_id + selection{mode:'query', filter:<FilterExpr>, exclude_ids} or {mode:'explicit', ids}). cell_filter picks which cells: 'empty_or_stale' (default in the UI; skips results still fresh under the current column config), 'failed', or 'all'; n_rows/start_row window the run (1-based). Returns lead_count + estimated_cost_usd + scope_resolved. ALWAYS show this to the user before run_table_column; it spends nothing.",
2188
2472
  inputSchema: obj(
2189
2473
  { list_id: S, column_id: S, view_id: S, selection: O, cell_filter: S, n_rows: N, start_row: N, only_failed: B },
2190
2474
  ['list_id', 'column_id'],
@@ -2198,7 +2482,7 @@ const TOOLS = {
2198
2482
  run_table_column: {
2199
2483
  def: {
2200
2484
  description:
2201
- "SPENDS vendor credits enrich one column across the scoped rows (view_id + selection + cell_filter + n_rows/start_row, same shape as estimate_table_column). cell_filter='empty_or_stale' re-runs only rows that are empty or out of date fresh results are never re-billed. Run estimate_table_column first and confirm with the user.",
2485
+ "SPENDS vendor credits: enrich one column across the scoped rows (view_id + selection + cell_filter + n_rows/start_row, same shape as estimate_table_column). cell_filter='empty_or_stale' re-runs only rows that are empty or out of date; fresh results are never re-billed. Run estimate_table_column first and confirm with the user.",
2202
2486
  inputSchema: obj(
2203
2487
  { list_id: S, column_id: S, view_id: S, selection: O, cell_filter: S, n_rows: N, start_row: N, only_failed: B },
2204
2488
  ['list_id', 'column_id'],
@@ -2212,7 +2496,7 @@ const TOOLS = {
2212
2496
  run_scope_summary: {
2213
2497
  def: {
2214
2498
  description:
2215
- 'FREE counts + estimates for a column run scope total, empty_or_stale, stale, failed rows + est_all/est_empty_or_stale/est_first_10 USD + budget. Use to decide what to run (which cell_filter) before estimate/run_table_column. Read-only.',
2499
+ 'FREE counts + estimates for a column run scope: total, empty_or_stale, stale, failed rows + est_all/est_empty_or_stale/est_first_10 USD + budget. Use to decide what to run (which cell_filter) before estimate/run_table_column. Read-only.',
2216
2500
  inputSchema: obj({ list_id: S, column_id: S, view_id: S, selection: O }, ['list_id', 'column_id']),
2217
2501
  },
2218
2502
  run: (a) =>
@@ -2226,7 +2510,7 @@ const TOOLS = {
2226
2510
  estimate_table_run_all: {
2227
2511
  def: {
2228
2512
  description:
2229
- 'FREE per-column preview + total for running EVERY enrichment column in dependency order, optionally scoped (view_id + selection). Show before run_table_all spends nothing.',
2513
+ 'FREE per-column preview + total for running EVERY enrichment column in dependency order, optionally scoped (view_id + selection). Show before run_table_all; spends nothing.',
2230
2514
  inputSchema: obj(
2231
2515
  { list_id: S, column_ids: ARR(S), view_id: S, selection: O, only_empty: B },
2232
2516
  ['list_id'],
@@ -2245,7 +2529,7 @@ const TOOLS = {
2245
2529
  run_table_all: {
2246
2530
  def: {
2247
2531
  description:
2248
- 'SPENDS vendor credits run every enrichment column in dependency order over the scoped rows (only empty cells run by default; cached results are free). Run estimate_table_run_all first and confirm with the user.',
2532
+ 'SPENDS vendor credits: run every enrichment column in dependency order over the scoped rows (only empty cells run by default; cached results are free). Run estimate_table_run_all first and confirm with the user.',
2249
2533
  inputSchema: obj(
2250
2534
  { list_id: S, column_ids: ARR(S), view_id: S, selection: O, only_empty: B },
2251
2535
  ['list_id'],
@@ -2275,7 +2559,7 @@ const TOOLS = {
2275
2559
  remove_table_rows: {
2276
2560
  def: {
2277
2561
  description:
2278
- 'Remove rows from a Scrapeloop table (list) by lead id the leads stay in the workspace pool; only the table membership and that table\'s enrichment cells are removed. Get ids from list_table_rows. To delete the whole table use delete_list.',
2562
+ 'Remove rows from a Scrapeloop table (list) by lead id. The leads stay in the workspace pool; only the table membership and that table\'s enrichment cells are removed. Get ids from list_table_rows. To delete the whole table use delete_list.',
2279
2563
  inputSchema: obj({ list_id: S, lead_ids: ARR(S) }, ['list_id', 'lead_ids']),
2280
2564
  },
2281
2565
  run: (a) =>
@@ -2364,7 +2648,7 @@ const TOOLS = {
2364
2648
  delete_list: {
2365
2649
  def: {
2366
2650
  description:
2367
- 'Delete a list (the list + its membership; the underlying leads stay in the workspace). Works for any Scrapeloop list. Irreversible confirm with the user first.',
2651
+ 'Delete a list (the list + its membership; the underlying leads stay in the workspace). Works for any Scrapeloop list. Irreversible: confirm with the user first.',
2368
2652
  inputSchema: obj({ list_id: S }, ['list_id']),
2369
2653
  },
2370
2654
  run: (a) =>
@@ -2375,7 +2659,7 @@ const TOOLS = {
2375
2659
  delete_campaign: {
2376
2660
  def: {
2377
2661
  description:
2378
- 'Delete a campaign (the Scrapeloop campaign + its ledger; the Instantly campaign itself is not deleted). Irreversible if you only want to stop sending, pause_campaign instead. Confirm with the user first.',
2662
+ 'Delete a campaign (the Scrapeloop campaign + its ledger; the Instantly campaign itself is not deleted). Irreversible. If you only want to stop sending, pause_campaign instead. Confirm with the user first.',
2379
2663
  inputSchema: obj({ campaign_id: S }, ['campaign_id']),
2380
2664
  },
2381
2665
  run: (a) =>
@@ -2388,7 +2672,7 @@ const TOOLS = {
2388
2672
  list_replies: {
2389
2673
  def: {
2390
2674
  description:
2391
- 'List inbound campaign replies, newest first. Filters: sentiment (interested|not_interested|unsubscribed), handled (true = already actioned, false = needs attention), lead_id, campaign_id, status (pending|auto_classified|human_required|human_reviewed human_required is the review queue).',
2675
+ 'List inbound campaign replies, newest first. Filters: sentiment (interested|not_interested|unsubscribed), handled (true = already actioned, false = needs attention), lead_id, campaign_id, status (pending|auto_classified|human_required|human_reviewed; human_required is the review queue).',
2392
2676
  inputSchema: obj({ sentiment: S, handled: B, lead_id: S, campaign_id: S, status: S, limit: N, offset: N }),
2393
2677
  },
2394
2678
  run: (a) =>
@@ -2408,7 +2692,7 @@ const TOOLS = {
2408
2692
  update_reply: {
2409
2693
  def: {
2410
2694
  description:
2411
- "Update a reply: mark it handled/unhandled and/or set its sentiment. Setting sentiment is a human classification it propagates to the lead (reply_sentiment, replied_at, status transition, the unsubscribe guard for 'unsubscribed', and the workspace's auto-suppress feed).",
2695
+ "Update a reply: mark it handled/unhandled and/or set its sentiment. Setting sentiment is a human classification: it propagates to the lead (reply_sentiment, replied_at, status transition, the unsubscribe guard for 'unsubscribed', and the workspace's auto-suppress feed).",
2412
2696
  inputSchema: obj(
2413
2697
  {
2414
2698
  reply_id: S,
@@ -2519,7 +2803,7 @@ const TOOLS = {
2519
2803
  {
2520
2804
  slug: S,
2521
2805
  status: { ...S, enum: ['planned', 'in_progress', 'deployed'] },
2522
- pr_url: { ...S, description: 'PR link set it when marking deployed' },
2806
+ pr_url: { ...S, description: 'PR link: set it when marking deployed' },
2523
2807
  },
2524
2808
  ['slug', 'status'],
2525
2809
  ),
@@ -2569,7 +2853,7 @@ const TOOLS = {
2569
2853
  roadmap_update_plan: {
2570
2854
  def: {
2571
2855
  description:
2572
- "Update a roadmap card's fields most importantly plan_md (the execution plan). Also: title, summary, kind, track, area, effort, depends_on, files_hint, position, pr_url.",
2856
+ "Update a roadmap card's fields, most importantly plan_md (the execution plan). Also: title, summary, kind, track, area, effort, depends_on, files_hint, position, pr_url.",
2573
2857
  inputSchema: obj(
2574
2858
  {
2575
2859
  slug: S,
@@ -2595,11 +2879,13 @@ const TOOLS = {
2595
2879
  );
2596
2880
  return Object.keys(body).length
2597
2881
  ? api('PATCH', `/roadmap/${enc(slug)}`, body)
2598
- : { ok: false, status: 400, error: { detail: 'nothing to update pass at least one field' } };
2882
+ : { ok: false, status: 400, error: { detail: 'nothing to update: pass at least one field' } };
2599
2883
  },
2600
2884
  },
2601
2885
  };
2602
2886
 
2887
+ TOOLS.preview_pipeline.def.inputSchema = TOOLS.create_pipeline.def.inputSchema;
2888
+
2603
2889
  const extractHttpMethods = (run) => [
2604
2890
  ...run.toString().matchAll(/api\(\s*['"](GET|POST|PUT|PATCH|DELETE)['"]/g),
2605
2891
  ].map((match) => match[1]);
@@ -2644,19 +2930,19 @@ for (const [name, tool] of Object.entries(TOOLS)) {
2644
2930
  }
2645
2931
 
2646
2932
  async function serve() {
2647
- // Start cleanly even without a key never crash or hang. The first tools/call
2933
+ // Start cleanly even without a key: never crash or hang. The first tools/call
2648
2934
  // returns a clear, structured auth error (api() handles the missing key), and we
2649
2935
  // log exactly one warning to stderr here. The key itself is never logged.
2650
2936
  if (!API_KEY) {
2651
2937
  warnStartupOnce(
2652
2938
  'missing_api_key',
2653
- 'scrapeloop-mcp: SCRAPELOOP_API_KEY is not set tools will return an auth error until it is. ' +
2939
+ 'scrapeloop-mcp: SCRAPELOOP_API_KEY is not set, so tools will return an auth error until it is. ' +
2654
2940
  'Generate a key in Scrapeloop → Settings → API access.'
2655
2941
  );
2656
2942
  }
2657
2943
 
2658
2944
  const server = new Server(
2659
- { name: 'scrapeloop-mcp', version: '0.7.1' },
2945
+ { name: 'scrapeloop-mcp', version: '0.8.0' },
2660
2946
  { capabilities: { tools: {} } }
2661
2947
  );
2662
2948
 
@@ -2697,8 +2983,8 @@ async function serve() {
2697
2983
  // Exported for the manifest/tools drift test. Tests set SCRAPELOOP_MCP_NO_SERVE
2698
2984
  // before importing so the stdio server (and the API-key requirement) stay dormant;
2699
2985
  // the CLI leaves it unset and serves. (An entrypoint-URL check is unreliable when
2700
- // the install path contains spaces import.meta.url percent-encodes them but
2701
- // process.argv[1] does not so an explicit opt-out flag is used instead.)
2986
+ // the install path contains spaces: import.meta.url percent-encodes them but
2987
+ // process.argv[1] does not, so an explicit opt-out flag is used instead.)
2702
2988
  export {
2703
2989
  MUTATION_POLICIES,
2704
2990
  TOOLS,
@@ -40,9 +40,14 @@ export const MUTATION_POLICIES = Object.freeze({
40
40
  create_integration: keyed('POST', ['list_integrations'], ['vendor', 'plugin_kind']),
41
41
  add_credential: reconcile('POST', ['list_credentials'], ['integration_id', 'label']),
42
42
  test_credential: idempotent('POST', ['list_credentials'], ['credential_id']),
43
+ preview_pipeline: read(),
44
+ preview_scrape_plan: read(),
45
+ create_scrape_plan: keyed('POST', ['list_scrape_plans'], ['name', 'target_list_id', 'kind']),
46
+ pause_scrape_plan: never('POST', ['get_scrape_plan'], ['plan_id']),
43
47
  estimate_scrape: read(),
44
48
  submit_scrape: keyed('POST', ['list_jobs'], ['kind', 'integration_id', 'credential_id', 'table_id']),
45
49
  cancel_job: never('POST', ['get_job'], ['job_id', 'status']),
50
+ attach_job_to_table: idempotent('POST', ['get_job'], ['job_id']),
46
51
  search_leads: read(),
47
52
  estimate_leads: read(),
48
53
  expand_job_titles: read(),
@@ -65,6 +70,8 @@ export const MUTATION_POLICIES = Object.freeze({
65
70
  run_enrich: never('POST', ['list_jobs'], ['preset_slug', 'preset_id', 'lead_ids', 'list_id']),
66
71
  test_recipe: never('POST', ['get_credits'], ['lead_id']),
67
72
  update_ai_settings: idempotent('PUT', ['get_ai_settings'], ['default_ai_model', 'enabled_models']),
73
+ update_playbook: idempotent('PUT', ['get_playbook'], ['channels', 'scrape', 'tables', 'checks', 'routing', 'campaigns']),
74
+ set_playbook_niches: idempotent('PUT', ['get_playbook'], ['niches']),
68
75
  update_ai_context: idempotent('PUT', ['get_ai_context'], ['company_description', 'icp']),
69
76
  generate_ai_context: never('POST', ['get_ai_context', 'get_credits'], ['domain']),
70
77
  create_strategy: reconcile('POST', ['list_strategies'], ['name']),
@@ -82,8 +89,19 @@ export const MUTATION_POLICIES = Object.freeze({
82
89
  pause_campaign: reconcile('POST', ['list_campaigns'], ['campaign_id', 'status']),
83
90
  sync_campaign: idempotent('POST', ['campaign_stats'], ['campaign_id']),
84
91
  update_sender_config: idempotent('PUT', ['get_sender_config'], ['credential_id']),
92
+ // A full-document PUT: sending it twice leaves the campaign in the same state
93
+ // as sending it once, so a lost response is safe to retry.
94
+ create_sender_fields: idempotent(
95
+ 'POST', ['get_campaign_field_mapping'], ['campaign_id', 'names'],
96
+ ),
97
+ set_campaign_field_mapping: idempotent(
98
+ 'PUT',
99
+ ['get_campaign_field_mapping'],
100
+ ['campaign_id'],
101
+ ),
85
102
  create_list: keyed('POST', ['get_lists'], ['name']),
86
103
  add_list_members: idempotent('POST', ['get_list_rows'], ['list_id', 'global_person_ids']),
104
+ create_pipeline: idempotent('POST', ['get_pipeline'], ['operation_id']),
87
105
  create_workbook: reconcile('POST', ['list_workbooks'], ['name', 'list_ids']),
88
106
  add_table_to_workbook: reconcile(
89
107
  'POST',
@@ -123,6 +141,14 @@ export const MUTATION_POLICIES = Object.freeze({
123
141
  ['get_table_columns'],
124
142
  ['table_id', 'preset_slug', 'preset_id', 'label', 'operation_id'],
125
143
  ),
144
+ // High-value create: with auto_run it starts a credit-billed backfill, and API
145
+ // idempotency is per (list_id, client_operation_id), so a fresh key on retry
146
+ // would make a second column AND a second paid backfill. One durable key.
147
+ create_table_native_column: keyed(
148
+ 'POST',
149
+ ['get_table_columns', 'get_table_native_tools'],
150
+ ['table_id', 'tool', 'label', 'operation_id'],
151
+ ),
126
152
  set_table_cells: idempotent('POST', ['list_table_rows'], ['table_id']),
127
153
  import_leads: idempotent(
128
154
  'POST',