Language router list

Response

200 - application/json

routers

object[]

routers.enabled

boolean

required

Is router enabled?

routers.models

object[]

required

the list of models that could handle requests

routers.models.anthropic

object

routers.models.anthropic.baseUrl

string

required

routers.models.anthropic.chatEndpoint

string

required

routers.models.anthropic.model

string

required

routers.models.anthropic.defaultParams

object

routers.models.anthropic.defaultParams.max_tokens

integer

routers.models.anthropic.defaultParams.metadata

string

routers.models.anthropic.defaultParams.stop

string[]

routers.models.anthropic.defaultParams.system

string

routers.models.anthropic.defaultParams.temperature

number

routers.models.anthropic.defaultParams.top_k

integer

routers.models.anthropic.defaultParams.top_p

number

routers.models.azureopenai

object

routers.models.azureopenai.apiVersion

string

required

The API version to use for this operation. This follows the YYYY-MM-DD format (e.g 2023-05-15)

routers.models.azureopenai.baseUrl

string

required

The name of your Azure OpenAI Resource (e.g https://glide-test.openai.azure.com/)

routers.models.azureopenai.model

string

required

This is your deployment name. You're required to first deploy a model before you can make calls (e.g. glide-gpt-35)

routers.models.azureopenai.chatEndpoint

string

routers.models.azureopenai.defaultParams

object

routers.models.azureopenai.defaultParams.frequency_penalty

integer

routers.models.azureopenai.defaultParams.logit_bias

object

routers.models.azureopenai.defaultParams.max_tokens

integer

routers.models.azureopenai.defaultParams.n

integer

routers.models.azureopenai.defaultParams.presence_penalty

integer

routers.models.azureopenai.defaultParams.response_format

object

routers.models.azureopenai.defaultParams.seed

integer

routers.models.azureopenai.defaultParams.stop

string[]

routers.models.azureopenai.defaultParams.temperature

number

routers.models.azureopenai.defaultParams.tool_choice

object

routers.models.azureopenai.defaultParams.tools

string[]

routers.models.azureopenai.defaultParams.top_p

number

routers.models.azureopenai.defaultParams.user

string

routers.models.client

object

routers.models.client.timeout

string

routers.models.cohere

object

routers.models.cohere.baseUrl

string

required

routers.models.cohere.chatEndpoint

string

required

routers.models.cohere.model

string

required

routers.models.cohere.defaultParams

object

routers.models.cohere.defaultParams.chat_history

object[]

routers.models.cohere.defaultParams.citiation_quality

string

routers.models.cohere.defaultParams.connectors

string[]

routers.models.cohere.defaultParams.conversation_id

string

routers.models.cohere.defaultParams.preamble_override

string

routers.models.cohere.defaultParams.prompt_truncation

string

routers.models.cohere.defaultParams.search_queries_only

boolean

routers.models.cohere.defaultParams.stream

boolean

unsupported right now

routers.models.cohere.defaultParams.temperature

number

routers.models.enabled

boolean

required

Is the model enabled?

routers.models.error_budget

string

routers.models.id

string

required

Model instance ID (unique in scope of the router)

routers.models.latency

object

routers.models.latency.decay

number

Weight of new latency measurements

routers.models.latency.update_interval

string

How often gateway should probe models with not the lowest response latency

routers.models.latency.warmup_samples

integer

The number of latency probes required to init moving average

routers.models.octoml

object

routers.models.octoml.baseUrl

string

required

routers.models.octoml.chatEndpoint

string

required

routers.models.octoml.model

string

required

routers.models.octoml.defaultParams

object

routers.models.octoml.defaultParams.frequency_penalty

integer

routers.models.octoml.defaultParams.max_tokens

integer

routers.models.octoml.defaultParams.presence_penalty

integer

routers.models.octoml.defaultParams.stop

string[]

routers.models.octoml.defaultParams.temperature

number

routers.models.octoml.defaultParams.top_p

number

routers.models.openai

object

Add other providers like

routers.models.openai.baseUrl

string

required

routers.models.openai.chatEndpoint

string

required

routers.models.openai.model

string

required

routers.models.openai.defaultParams

object

routers.models.openai.defaultParams.frequency_penalty

integer

routers.models.openai.defaultParams.logit_bias

object

routers.models.openai.defaultParams.max_tokens

integer

routers.models.openai.defaultParams.n

integer

routers.models.openai.defaultParams.presence_penalty

integer

routers.models.openai.defaultParams.response_format

object

routers.models.openai.defaultParams.seed

integer

routers.models.openai.defaultParams.stop

string[]

routers.models.openai.defaultParams.temperature

number

routers.models.openai.defaultParams.tool_choice

object

routers.models.openai.defaultParams.tools

string[]

routers.models.openai.defaultParams.top_p

number

routers.models.openai.defaultParams.user

string

routers.models.weight

integer

routers.retry

object

required

retry when no healthy model is available to router

routers.retry.base_multiplier

integer

routers.retry.max_delay

integer

routers.retry.max_retries

integer

routers.retry.min_delay

integer

routers.routers

string

required

Unique router ID

routers.strategy

string

required

strategy on picking the next model to serve the request

API Documentation

Response