Skip to main content
POST
/
langchain
/
cohere
/
v2
/
chat
Error
A valid request URL is required to generate request examples
{
  "id": "<string>",
  "message": {
    "role": "<string>",
    "content": [
      {
        "text": "<string>",
        "image_url": {
          "url": "<string>"
        },
        "thinking": "<string>",
        "document": {
          "data": {},
          "id": "<string>"
        }
      }
    ],
    "tool_calls": [
      {
        "id": "<string>",
        "function": {
          "name": "<string>",
          "arguments": "<string>"
        }
      }
    ],
    "tool_plan": "<string>"
  },
  "usage": {
    "billed_units": {
      "input_tokens": 123,
      "output_tokens": 123,
      "search_units": 123,
      "classifications": 123
    },
    "tokens": {
      "input_tokens": 123,
      "output_tokens": 123
    },
    "cached_tokens": 123
  },
  "logprobs": [
    {
      "token_ids": [
        123
      ],
      "text": "<string>",
      "logprobs": [
        123
      ]
    }
  ]
}

Documentation Index

Fetch the complete documentation index at: https://docs.getbifrost.ai/llms.txt

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

Authorizations

Authorization
string
header
required

Bearer token authentication. Use your provider API key or Bifrost authentication token. Virtual keys (prefixed with sk-bf-) can also be passed here.

Body

application/json
model
string
required

Model to use for chat completion

Example:

"command-r-plus"

messages
object[]
required

Array of message objects

tools
object[]
tool_choice
enum<string>

Tool choice mode - AUTO lets the model decide, NONE disables tools, REQUIRED forces tool use

Available options:
AUTO,
NONE,
REQUIRED
temperature
number
Required range: 0 <= x <= 1
p
number

Top-p sampling

k
integer

Top-k sampling

max_tokens
integer
stop_sequences
string[]
frequency_penalty
number
presence_penalty
number
stream
boolean
safety_mode
enum<string>
Available options:
CONTEXTUAL,
STRICT,
NONE
log_probs
boolean
strict_tool_choice
boolean
thinking
object
response_format
object

Response

Successful response

id
string
finish_reason
enum<string>
Available options:
COMPLETE,
STOP_SEQUENCE,
MAX_TOKENS,
TOOL_CALL,
ERROR,
TIMEOUT
message
object
usage
object
logprobs
object[]

Log probabilities (if requested)