packages/ai-cohere
| Name | Last commit message | Last commit date |
|---|---|---|
| .. | ||
README.md
@tanstack/ai-cohere#
Cohere adapter for TanStack AI. It gives you two
things: multimodal embeddings with embed-v4.0, and document reranking with
Cohere's rerank models — the precision step for RAG and search pipelines.
For chat, summarization, or media, use OpenAI, Anthropic, or Gemini.
Install#
pnpm add @tanstack/ai @tanstack/ai-cohere
Setup#
Get your API key from the Cohere Dashboard and set it as an environment variable:
export COHERE_API_KEY="..."
Usage#
Embedding Adapter#
import { cohereEmbedding } from '@tanstack/ai-cohere'
import { embed } from '@tanstack/ai'
const adapter = cohereEmbedding('embed-v4.0')
const result = await embed({
adapter,
input: ['a red guitar', 'a blue drum kit'],
modelOptions: { inputType: 'search_document' },
})
console.log(result.embeddings[0].vector)
Multimodal Inputs#
embed-v4.0 embeds text, images, and fused text+image items (one vector per input item). Fuse parts by nesting them in an array — the outer array is the item list:
const result = await embed({
adapter,
input: [
'a red guitar',
{
type: 'image',
source: { type: 'data', value: base64Png, mimeType: 'image/png' },
},
// A nested array fuses its parts into a single vector.
[
{ type: 'text', content: 'product photo' },
{
type: 'image',
source: { type: 'data', value: base64Jpeg, mimeType: 'image/jpeg' },
},
],
],
modelOptions: { inputType: 'search_document' },
})
Cohere does not fetch remote image URLs. Pass base64 data or a data: URI, or enable allowUrlFetch in the adapter config to have the adapter download http(s) URLs and inline them.
Rerank Adapter#
import { rerank } from '@tanstack/ai'
import { cohereRerank } from '@tanstack/ai-cohere'
const { ranking, rerankedDocuments } = await rerank({
adapter: cohereRerank('rerank-v3.5'),
query: 'talk about rain',
documents: ['sunny day at the beach', 'rainy afternoon in the city'],
topN: 2,
})
console.log(rerankedDocuments[0]) // 'rainy afternoon in the city'
With Explicit API Key#
Both adapters read COHERE_API_KEY from the environment. To pass a key
explicitly, use the create* factories:
import { createCohereEmbedding, createCohereRerank } from '@tanstack/ai-cohere'
const embedAdapter = createCohereEmbedding(
'embed-v4.0',
process.env.COHERE_API_KEY!,
)
const rerankAdapter = createCohereRerank('rerank-v3.5', 'co-...')
Supported Models#
Embedding Models#
embed-v4.0- Multimodal embedding model (text + images, Matryoshka dimensions via the top-leveldimensionsoption)
Rerank Models#
rerank-v3.5rerank-english-v3.0rerank-multilingual-v3.0
Features#
- ✅ Embeddings (batch, one request per input array)
- ✅ Multimodal embedding input (text + images + fused text/image items)
- ✅ Dimension reduction (
dimensions→ Cohereoutput_dimension) - ✅ Document reranking
- ❌ Chat / text generation
- ❌ Image generation
Tree-Shakeable Adapters#
This package uses tree-shakeable adapters, so you only import what you need:
import { cohereEmbedding, cohereRerank } from '@tanstack/ai-cohere'
Read the docs ->#
- Reranking Guide — object documents, RAG pipelines, options, and the result shape.
- Cohere Adapter — models, configuration, and explicit API keys.
License#
MIT