Crate anyllm_openai_compat 

Reusable OpenAI-compatible adapter toolkit for anyllm provider authors.

Use this crate when building a new anyllm provider crate for an upstream API that is close enough to OpenAI’s Chat Completions shape that wire translation, SSE parsing, and response conversion can be shared. It is not primarily intended for end-user application code; first-party provider crates (anyllm-openai, anyllm-cloudflare-worker) consume it.

See providers for pre-configured factories and the crate README for the intended adapter shape, observability notes, and auth contracts.

use anyllm_openai_compat::providers::Cloudflare;

let provider = Cloudflare::new("cf-account-id", "cf-api-token")?;

Modules

providers

Pre-configured provider factories for common OpenAI-compatible APIs.

Structs

ChatCompletionRequest

ChatCompletionResponse

EmbeddingData

EmbeddingRequestOptions

Wire-level extras applied to the OpenAI-style embedding request.

EmbeddingsRequest

OpenAI-style embeddings request wire shape.

EmbeddingsResponse

OpenAI-style embeddings response wire shape.

EmbeddingsUsage

Provider

A generic OpenAI-compatible chat provider.

ProviderBuilder

Builder for configuring a Provider.

RequestOptions

SseState

TransportConfig

Enums

OpenAIReasoningEffort

OpenAI-compatible reasoning effort levels.

Functions

extract_request_id

extract_retry_after

from_api_response

from_embeddings_response

Convert an OpenAI-style embeddings response into the portable EmbeddingResponse.

map_http_error

map_response_deserialize_error

map_stream_error

map_transport_error

parse_finish_reason

process_sse_data

send_chat_completion_request

send_embeddings_request

Generic HTTP dispatch for sending an OpenAI-style embeddings request body via a caller-supplied send future. Mirrors crate::send_chat_completion_request.

sse_to_stream

to_chat_completion_request

to_embeddings_request

Convert an anyllm EmbeddingRequest into the OpenAI-style wire shape.