@myapihq/cli 2.2.0 → 2.3.1

package/dist/commands/llm.js

@@ -47,6 +47,25 @@ async function defaultChatModel(apiKey, orgId) {
     }
     return chat.id;
 }
+// Resolve the embed model from --model, else the live catalog. Unlike chat
+// we don't silently pick a default when several are served — embedding
+// vectors aren't cross-model compatible, so an ambiguous pick is a footgun.
+// One served embed model → use it; none → catalog-truthful error (not a
+// hardcoded "not available today"); many → make the caller choose.
+async function resolveEmbedModel(apiKey, orgId, flags) {
+    if (typeof flags.model === 'string' && flags.model)
+        return flags.model;
+    const embeds = (await loadModels(apiKey, orgId)).filter(m => m.kind === 'embed');
+    if (embeds.length === 1)
+        return embeds[0].id;
+    if (embeds.length === 0) {
+        error('No embedding model is currently served — the catalog has none. Run "myapi llm models --kind embed" to check.');
+    }
+    else {
+        error(`Multiple embedding models available — pass --model <id>. Options: ${embeds.map(m => m.id).join(', ')}`);
+    }
+    throw new Error('unreachable'); // satisfy the type checker
+}
 function readPromptArg(promptArg, flags) {
     if (typeof flags.file === 'string' && flags.file) {
         return fs.readFileSync(flags.file, 'utf-8');
@@ -140,10 +159,8 @@ async function complete(promptArg, flags) {
 }
 async function embed(inputArg, flags) {
     const config = requireConfig();
-    const orgId = requireOrg(flags, config, 'myapi llm embed "<text>" --model <id> [--org <id>]');
-    const model = typeof flags.model === 'string' ? flags.model : '';
-    if (!model)
-        error('Missing required flag: --model. Run "myapi llm models" to list embed models.');
+    const orgId = requireOrg(flags, config, 'myapi llm embed "<text>" [--model <id>] [--org <id>]');
+    const model = await resolveEmbedModel(config.api_key, orgId, flags);
     const text = readPromptArg(inputArg, flags);
     if (!text.trim())
         error('Empty input. Pass <text> as an argument, "-" to read stdin, or --file <path>.');
@@ -283,8 +300,7 @@ const SUBCOMMAND_USAGE = {
        [--stop <csv>] [--file <path>] [--org <id>] [--json]
 
   Raw chat completion against a self-hosted catalog model. Without --model,
-  picks the first chat model from "myapi llm models" (today:
-  Qwen/Qwen3-Coder-30B-A3B-Instruct).
+  picks the first chat model from "myapi llm models".
 
   Pass "-" as the prompt to read from stdin. The reply goes to stdout; a
   one-line usage footer (tokens + cost in cents) goes to stderr so it
@@ -293,10 +309,11 @@ const SUBCOMMAND_USAGE = {
   Examples:
     myapi llm complete "summarize: $(cat README.md)"
     cat draft.md | myapi llm complete - --system "You are an editor" --max-tokens 200`,
-    embed: `myapi llm embed "<text>" --model <id> [--file <path>] [--org <id>] [--json]
+    embed: `myapi llm embed "<text>" [--model <id>] [--file <path>] [--org <id>] [--json]
 
-  No embedding model is served on the raw catalog today — this returns
-  EMBED_NOT_AVAILABLE until one is. Use a dedicated embedding API for now.`,
+  Embed text into a dense vector. --model is optional when the catalog serves
+  exactly one embedding model; otherwise pass one from "myapi llm models
+  --kind embed". Returns EMBED_NOT_AVAILABLE if no embedding model is served.`,
     models: `myapi llm models [--kind chat|embed] [--org <id>] [--json]
 
   Lists the live model catalog with per-1M-token pricing in cents.
@@ -358,7 +375,7 @@ Subcommands:
   classify    Pick a label from a set
   complete    Raw chat completion against a catalog model
   draft       Write something (email | reply | message | …)
-  embed       Raw embeddings (no model served on raw today)
+  embed       Raw embeddings against a catalog embed model
   extract     Pull structured data conforming to a JSON Schema
   models      List the live model catalog (id, kind, context, cents/1M)
   summarize   Summarize text (brief | exec | bullet)

package/dist/skills/my-llm-api/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: my-llm-api
-version: 1.1.0
+version: 1.2.0
 description: >
   Two-surface LLM primitive. Raw chat completion against self-hosted
   open-source models (you pick the model), and objective verbs
@@ -14,7 +14,7 @@ checksum: sha256-pending
 
 A two-surface LLM gateway:
 
-- **Raw** — `complete` / `embed` / `models`. You pick a self-hosted catalog model (today: `Qwen/Qwen3-Coder-30B-A3B-Instruct`). Proprietary models are not callable here.
+- **Raw** — `complete` / `embed` / `models`. You pick a self-hosted catalog model (run `myapi llm models` for the live list). Proprietary models are not callable here.
 - **Verbs** — `classify` / `extract` / `summarize` / `draft`. You ask for a task done; the model is implementation detail and is never named in the response.
 
 Pricing is in cents per 1M tokens at the actual upstream rate; cost is debited from your MyAPI balance.
@@ -33,7 +33,7 @@ Use this for workflow tasks — summarize a doc, classify an inbound email, extr
 ### Raw `complete` request
 ```json
 {
-  "model": "Qwen/Qwen3-Coder-30B-A3B-Instruct",
+  "model": "<model-id>",
   "messages": [
     { "role": "system", "content": "You are a terse editor." },
     { "role": "user",   "content": "Tighten this paragraph: ..." }
@@ -49,7 +49,7 @@ Roles: `system | user | assistant`. Multiple system messages collapse to one ins
 ### Raw `complete` response
 ```json
 {
-  "model": "Qwen/Qwen3-Coder-30B-A3B-Instruct",
+  "model": "<model-id>",
   "content": "...assistant reply...",
   "finish_reason": "stop",
   "usage": { "input_tokens": 42, "output_tokens": 87, "cost_cents": 0.029 }
@@ -62,7 +62,7 @@ If `model` isn't in the self-hosted catalog the server returns `MODEL_NOT_IN_RAW
 
 ### Raw `embed`
 
-No embedding model is served on the raw catalog today; calls return `EMBED_NOT_AVAILABLE`. Use a dedicated embedding API for now.
+Embed text into a dense vector. `--model` is optional when the catalog serves exactly one embed model; else pass one from `myapi llm models --kind embed`. Returns `EMBED_NOT_AVAILABLE` when none is served.
 
 ### Model catalog
 - Chat models: `id`, `kind: 'chat'`, `context_window`, `input_cost_per_1m_cents`, `output_cost_per_1m_cents`
@@ -99,11 +99,11 @@ client = OpenAI(
     api_key="hq_live_…",
     base_url="https://api.myapihq.com/llm/orgs/<org_id>/v1",
 )
-r = client.chat.completions.create(model="Qwen/Qwen3-Coder-30B-A3B-Instruct",
+r = client.chat.completions.create(model="<model-id>",
     messages=[{"role":"user","content":"Hi"}])
 ```
 
-Use raw `complete` (MyAPI shape) for first-party integrations; use the OpenAI-compat path for compatibility with existing client code.
+Use raw `complete` for first-party code; the compat path for existing OpenAI/LangChain tooling.
 
 <!-- llm:end -->
 
@@ -113,7 +113,7 @@ Use raw `complete` (MyAPI shape) for first-party integrations; use the OpenAI-co
 |---|---|
 | `myapi llm models [--kind chat\|embed] [--json]` | List the live model catalog with pricing (cents/1M) |
 | `myapi llm complete "<prompt>" [--model <id>] [--system "<s>"] [--max-tokens N] [--temperature 0..1] [--stop <csv>] [--file <path>] [--json]` | Raw chat completion; reply to stdout, usage to stderr. Defaults to the first chat model in the catalog |
-| `myapi llm embed "<text>" --model <id> [--json]` | Embed a string (no model served today — returns EMBED_NOT_AVAILABLE) |
+| `myapi llm embed "<text>" [--model <id>] [--json]` | Embed a string into a vector; `--model` optional when the catalog has one embed model |
 | `myapi llm classify "<input>" --labels <csv> [--multi] [--tier <t>] [--json]` | Pick a label from a set |
 | `myapi llm extract "<input>" --schema <path\|json> [--tier <t>] [--json]` | Pull structured data conforming to a JSON Schema |
 | `myapi llm summarize "<input>" [--style brief\|exec\|bullet] [--tier <t>] [--json]` | Summarize text |
@@ -132,9 +132,9 @@ myapi llm models --kind chat --json | jq '.models[].id'
 # Raw completion — picks the first chat model from the catalog
 myapi llm complete "Summarize in 12 words: $(cat README.md)"
 
-# Pin a specific model
+# Pin a specific model (ids come from `myapi llm models`)
 myapi llm complete "Refactor this function: ..." \
-  --model Qwen/Qwen3-Coder-30B-A3B-Instruct \
+  --model <model-id> \
   --system "You are a careful Go reviewer." \
   --max-tokens 600
 
@@ -163,5 +163,5 @@ INTENT=$(printf '%s' "$BODY" | myapi llm classify - \
 
 - **`draft` context safety.** `context` fields are quoted into the prompt verbatim; sensitive-named keys (`secret`, `api_key`, `password`, …) are NOT redacted. Two guards on top: (a) injection-defense strips `instructions`/`system`/`prompt`/`override` keys and surfaces them in `meta.warnings`; (b) output guardrail substring-scans fact values (length ≥ 4) in the response and lists matches in `meta.guardrails.facts_in_output` (signal, not redaction). Rule of thumb: never put credentials, PII, or internal metadata in `context` — pass identifiers, reference them indirectly.
 - **Self-hosted raw, server-picked verbs.** Raw runs on MyAPI's TPU; verbs route wherever the server picks.
-- **Cost + latency.** `usage.cost_cents` is authoritative — no markup. Qwen3-Coder-30B: 200–600 ms to first token, 1–3 s end-to-end.
+- **Cost + latency.** `usage.cost_cents` is authoritative — no markup. Varies by tier: 200–600 ms to first token, 1–3 s end-to-end.
 - **Live catalog, no streaming, no BYOK.** Don't hard-code ids — `models` is truth (CLI auto-picks if `--model` omitted). Full reply only.

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@myapihq/cli",
   "license": "Apache-2.0",
-  "version": "2.2.0",
+  "version": "2.3.1",
   "description": "MyAPI command-line interface",
   "type": "module",
   "files": [
@@ -31,7 +31,7 @@
     "lint:skills:strict": "node scripts/copy-skills.js && node scripts/lint-skills.js --strict"
   },
   "dependencies": {
-    "@myapihq/sdk": "^2.2.0"
+    "@myapihq/sdk": "^2.3.1"
   },
   "devDependencies": {
     "@types/node": "^25.6.0",