Cohere

Cohere integration (AI). Imported from open-connector.

API key3 toolshttps://api.cohere.com

POST

cohere_chat

Generate a synchronous text response using the Cohere Chat API.

/v2/chat

20 parameters

Param	Type	Required	Description
model	string	required	The name of a compatible Cohere chat model.
messages	array	required	The ordered conversation messages to send to Cohere.
tools	array	optional	Function tools available to the model.
strict_tools	boolean	optional	Whether tool calls must strictly follow the tool definition.
documents	array	optional	Documents that the model can cite while generating the response.
citation_options	object	optional	A JSON object accepted or returned by the Cohere API.
response_format	object	optional	A JSON object accepted or returned by the Cohere API.
safety_mode	string	optional	The Cohere safety mode to use for the request.
max_tokens	integer	optional	The maximum number of output tokens the model will generate.
stop_sequences	array	optional	Stop sequences that halt generation.
temperature	number	optional	Randomness used for generation.
seed	integer	optional	A best-effort deterministic sampling seed.
frequency_penalty	number	optional	Penalty used to reduce repeated tokens.
presence_penalty	number	optional	Penalty used to reduce reused token content.
k	integer	optional	Top-k sampling value. Use 0 to disable k-sampling.
p	number	optional	Top-p sampling value.
logprobs	boolean	optional	Whether to include generated-token log probabilities.
tool_choice	string	optional	Tool selection behavior for the model.
thinking	object	optional	A JSON object accepted or returned by the Cohere API.
priority	integer	optional	Lower values are handled earlier by Cohere.

POST

cohere_embed_texts

Generate embeddings for text inputs using the Cohere Embed API.

/v2/embed

8 parameters

Param	Type	Required	Description
model	string	required	The Cohere embedding model identifier.
input_type	string	required	The type of input passed to the embedding model.
texts	array	required	Text strings for the model to embed.
max_tokens	integer	optional	The maximum number of tokens to embed per input.
output_dimension	integer	optional	The number of dimensions for embed-v4 and newer models.
embedding_types	array	optional	Embedding representations to return.
truncate	string	optional	How Cohere handles inputs longer than the model limit.
priority	integer	optional	Lower values are handled earlier by Cohere.

POST

cohere_rerank_documents

Rank text documents by relevance to a query using the Cohere Rerank API.

/v2/rerank

6 parameters

Param	Type	Required	Description
model	string	required	The Cohere rerank model identifier.
query	string	required	The search query used to rank documents.
documents	array	required	Texts that will be compared to the query.
top_n	integer	optional	Maximum number of rerank results to return.
max_tokens_per_doc	integer	optional	Maximum tokens to keep per document before ranking.
priority	integer	optional	Lower values are handled earlier by Cohere.