Skip to main content
POST
/
embeddings
/
text
Generate Text Embedding
curl --request POST \
  --url https://mavi-backend.memories.ai/serve/api/v2/embeddings/text \
  --header 'Authorization: <api-key>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "input": [
    "Sample text to embed",
    "Another text string"
  ],
  "model": "gemini-embedding-001",
  "dimensionality": 512
}
'
{
  "code": 200,
  "msg": "success",
  "data": [
    {
      "embedding": [
        -0.031649195,
        0.00069497403,
        0.011988669,
        "... (continues for `dimensionality` length)"
      ]
    },
    {
      "embedding": [
        0.012345,
        -0.067890,
        "... (one entry per input string)"
      ]
    }
  ],
  "success": true,
  "failed": false
}

Documentation Index

Fetch the complete documentation index at: https://api-tools.memories.ai/llms.txt

Use this file to discover all available pages before exploring further.

Product: Visual Intelligence — Embeddings Use case: Generate vector embeddings for image, video, or text inputs for semantic search and similarity tasks Host: https://mavi-backend.memories.ai/serve/api/v2 Auth: Authorization: sk-mavi-... (no Bearer prefix)
This endpoint generates vector embeddings for text strings. You can embed single or multiple text inputs in one request, and optionally specify the output dimensionality.
Pricing:
  • $0.0000002 per token (2e-7/token)

​
Code Examples

const BASE_URL = "https://mavi-backend.memories.ai/serve/api/v2";
const API_KEY = "sk-mavi-...";

const response = await fetch(`${BASE_URL}/embeddings/text`, {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': API_KEY
  },
  body: JSON.stringify({
    input: ['Sample text to embed', 'Another text string'],
    model: 'gemini-embedding-001',
    dimensionality: 512
  })
});

const data = await response.json();
console.log(data);

​
Request Body

FieldTypeRequiredDescription
inputarray[string]YesList of text strings to embed (can be a single string or multiple strings)
modelstringYesEmbedding model name
dimensionalityintegerNoOutput embedding dimension size (e.g., 256, 512, 768). Model-dependent.
Supported Models:
  • gemini-embedding-001 - Google’s Gemini text embedding model

​
Response

Returns embedding vectors for each input text.
The live response shape does not match the previously documented “singular vs plural” split. Live response is always data: [ {embedding: [...]} ] — an array of objects, each carrying its own embedding (singular) key — regardless of whether you sent one input or many. There is no top-level data.embedding / data.embeddings field. Verified live with gemini-embedding-001.
{
  "code": 200,
  "msg": "success",
  "data": [
    {
      "embedding": [
        -0.031649195,
        0.00069497403,
        0.011988669,
        "... (continues for `dimensionality` length)"
      ]
    },
    {
      "embedding": [
        0.012345,
        -0.067890,
        "... (one entry per input string)"
      ]
    }
  ],
  "success": true,
  "failed": false
}

​
Response Parameters

ParameterTypeDescription
codeintegerResponse code (200 on success)
msgstringResponse message describing the operation result
dataarray[object]One entry per input string, in the same order.
data[].embeddingarray[number]Vector for that input. Length equals dimensionality (or the model default).
successbooleanIndicates whether the operation was successful
failedbooleanIndicates whether the operation failed

​
Notes

  • Text embeddings are returned synchronously in the response.
  • The response is always an array — index into data[i].embedding to get the vector for input[i]. There is no singular-vs-plural variant.
  • The dimensionality parameter allows you to control the output vector size
  • Supported dimensionality depends on the model (common values: 256, 512, 768, 1024)
  • Lower dimensionality results in faster processing and reduced storage, but may have lower accuracy
  • Use text embeddings for:
    • Semantic search and similarity matching
    • Text classification and clustering
    • Question answering and information retrieval
    • Recommendation systems
    • Duplicate detection

Authorizations

​
Authorization
string
header
required

Body

application/json
​
input
string[]
required

List of text strings to embed

Example:
[
  "Sample text to embed",
  "Another text string"
]
​
model
enum<string>
required

Embedding model name

Available options:
gemini-embedding-001
Example:

"gemini-embedding-001"

​
dimensionality
integer

Output embedding dimension size

Example:

512

Response

200 - application/json

Embeddings generated successfully

Response for single input

​
code
string

Response code indicating the result status

Example:

200

​
msg
string

Response message describing the operation result

Example:

"success"

​
data
object

Response data object containing the embedding

​
success
boolean

Indicates whether the operation was successful

Example:

true

​
failed
boolean

Indicates whether the operation failed

Example:

false