openrouter-embeddings

Text → embedding vector via OpenRouter. Default model: qwen/qwen3-embedding-8b.

Usage

Resolve TOOL_DIR = the directory containing this SKILL.md. Commands below use TOOL_DIR as a symbolic placeholder; replace it with the resolved, quoted path before running Bash.

Single text

export OPENROUTER_API_KEY=sk-or-v1-...

python3 TOOL_DIR/scripts/embed.py \
  --text "The quick brown fox jumps over the lazy dog" \
  --output vec.json

Batch from JSONL

Input records.jsonl (one JSON per line):

{"id": "row_0", "text": "Every place name in the United States."}
{"id": "row_1", "text": "Nearby stars and potential exoplanets."}

Run:

python3 TOOL_DIR/scripts/embed.py \
  --jsonl records.jsonl \
  --output records_with_embeddings.jsonl \
  --batch-size 32

Output is the same JSONL with an added embedding field per line.

Flags

Flag	Default	Description
--text	—	Embed one string (mutually exclusive with --jsonl)
--jsonl	—	Embed many; each line must have a text field
--output	required	Output path
--model	qwen/qwen3-embedding-8b	Any embedding model on OpenRouter
--batch-size	32	Records per API call (jsonl mode)
--dimensions	—	Optional: truncate to N dims if supported

Endpoint

POST /api/v1/embeddings — OpenAI-compatible schema.

Request:

{ "model": "qwen/qwen3-embedding-8b", "input": ["text1", "text2", ...] }

Response:

{ "data": [ { "embedding": [0.01, -0.02, ...], "index": 0 }, ... ], "model": "...", "usage": {...} }

Notes

• qwen3-embedding-8b outputs high-dimensional dense vectors suitable for semantic similarity, clustering, RAG.
• For cheaper batches, consider qwen/qwen3-embedding-4b or other listed embedding models (GET /api/v1/embeddings/models).