Embedding

ragrails embed reads chunk JSON from --input-dir, generates vectors, and writes embedded.json to --output-dir.

pip install "ragrails[voyage]"
export VOYAGE_API_KEY="..."

Command

ragrails embed \
  --input-dir files/output/chunks/ \
  --output-dir files/output/embedded/ \
  --provider voyage \
  --model voyage-3 \
  --batch-size 64

Output:

Inputs   : 87
Embedded : 87
Failed   : 0
Saved to : files/output/embedded/embedded.json

The CLI embeds with document input type for indexing. Query-time embedding happens inside query, retrieve, and chat.

Use the same embedding provider/model for indexing and retrieval. Vectors from different models are not comparable.

Options

Option	Default	Description
`--input-dir`	required	Folder containing chunk JSON files.
`--output-dir`	required	Folder to write `embedded.json`.
`--provider`	`voyage`	Embedding provider.
`--model`	`voyage-3`	Embedding model.
`--batch-size`	`64`	Chunks per embedding request.

Next Stage

ragrails store --input-dir files/output/embedded/ \
  --vector-db qdrant --collection support --url http://localhost:6333

Chunking

Create chunks.

Storing

Store embedded chunks.

Get Started

Workflows

Pipeline

Maintenance

Command

Options

Next Stage

Chunking

Storing

​Command

​Options

​Next Stage

Chunking

Storing

Command

Options

Next Stage