Create Embeddings
jina.embed
ActionAI & modelsconnection required
Calls the Jina AI API with the connection's credentials.
Turn text, images or code into embedding vectors with a Jina embedding model, for semantic search or clustering.
Connection
- Required
- Yes — the node will not save without one
- Auth type
- Bearer token
- Providers
- Jina AI
- Secret shape
- the API key, or {"api_key": "…"}
- Config it needs
- base_url, default_embedding_model, default_reranker_model
- Resolved by
- config.connection_id, else node.connection_id
Configuration
8 fields, 1 required. A dotted path is a nested object and name[] an array of objects — the editor renders the same tree.
| Field | Type | Required | Default | Constraints | Example |
|---|---|---|---|---|---|
input Items to embed: strings, or objects such as {"text": "…"} and {"image": "https://…"} for the multimodal models. | array | Yes | — | — | ["the quick brown fox"] |
model Embedding model, for example jina-embeddings-v5-text-small or jina-embeddings-v4. Falls back to the connection's default. | string | No | — | — | jina-embeddings-v5-text-small |
task Downstream task the vectors are optimised for. It must be one the chosen model supports. | string | No | — | one of "retrieval.query", "retrieval.passage", "text-matching", "classification", "clustering", "separation", "code.query", "code.passage", "nl2code.query", "nl2code.passage", "code2code.query", "code2code.passage", "code2nl.query", "code2nl.passage", "code2completion.query", "code2completion.passage", "qa.query", "qa.passage" | — |
dimensions Truncate each vector to this many dimensions; the ceiling is the model's own output size. | number | No | — | 1–2048 | — |
embedding_type Encoding of the returned vectors. | string | No | float | one of "float", "base64", "binary", "ubinary" | — |
normalized Return L2-normalised unit vectors. | boolean | No | — | — | — |
late_chunking Embed the inputs as one long document, then split — keeps chunk vectors aware of the whole text. Not on every model. | boolean | No | — | — | — |
truncate Drop input past the model's context limit instead of failing. | boolean | No | — | — | — |
Example configuration
Every value below is the contract’s own example or default — nothing invented.
{
"embedding_type": "float",
"input": [
"the quick brown fox"
],
"model": "jina-embeddings-v5-text-small"
}Output
One shape.
result
| Field | Type | Example |
|---|---|---|
model The model that produced the vectors. | string | — |
embeddings One vector per input, in input order. | array | — |
data Jina's raw embedding objects, with their index. | array | — |
count How many vectors came back. | number | — |
usage Token usage Jina billed for this call. | object | — |
{
"count": 0,
"data": [],
"embeddings": [],
"model": "",
"usage": {}
}See Create Embeddings run in a workflow.
In a technical demo we add the Jina AI connection, drop the node on the canvas and let Workflow Health pre-flight it before the first real run.