# llm_call v0.1.0 - API Reference

## Modules

- [LlmCall](LlmCall.md): Wire-protocol codec for LLM APIs.
- [LlmCall.Application](LlmCall.Application.md): Starts the default Finch pool.
- [LlmCall.Capabilities](LlmCall.Capabilities.md): What a completion protocol supports.
- [LlmCall.Client](LlmCall.Client.md): A provider, resolved for calling.
- [LlmCall.Codec](LlmCall.Codec.md): The behaviour every dialect adapter implements.
- [LlmCall.Codec.Conformance](LlmCall.Codec.Conformance.md): Fixture-based assertions for codec authors.
- [LlmCall.Codec.Info](LlmCall.Codec.Info.md): Codec metadata.
- [LlmCall.Config](LlmCall.Config.md): Codec defaults and capture policy.
- [LlmCall.Conformance](LlmCall.Conformance.md): Assertions shared by `LlmCall.Codec.Conformance` and
`LlmCall.Media.Conformance`.
- [LlmCall.Content](LlmCall.Content.md): Canonical content parts.
- [LlmCall.Content.File](LlmCall.Content.File.md): A document or file block.
- [LlmCall.Content.Image](LlmCall.Content.Image.md): An image block. `source` is a URL, base64 data, or a provider file id.
- [LlmCall.Content.Raw](LlmCall.Content.Raw.md): A provider block the codec does not understand.
- [LlmCall.Content.Text](LlmCall.Content.Text.md): A text block.
- [LlmCall.Content.Thinking](LlmCall.Content.Thinking.md): A reasoning block.
- [LlmCall.Content.Video](LlmCall.Content.Video.md): A video block, used for references and generated assets.
- [LlmCall.Dialect.Adapter](LlmCall.Dialect.Adapter.md): `use` for a completion adapter.
- [LlmCall.Dialect.Anthropic.Messages](LlmCall.Dialect.Anthropic.Messages.md): Anthropic Messages (`POST /v1/messages`).
- [LlmCall.Dialect.Gemini.GenerateContent](LlmCall.Dialect.Gemini.GenerateContent.md): Gemini `generateContent` / `streamGenerateContent` (`v1beta`).
- [LlmCall.Dialect.Helpers](LlmCall.Dialect.Helpers.md): Validation and encoding primitives shared by dialect adapters.
- [LlmCall.Dialect.OpenAI.Chat](LlmCall.Dialect.OpenAI.Chat.md): OpenAI Chat Completions (`POST /v1/chat/completions`).
- [LlmCall.Dialect.OpenAI.Responses](LlmCall.Dialect.OpenAI.Responses.md): OpenAI Responses (`POST /v1/responses`).
- [LlmCall.Dialect.OpenAI.Responses.Connection](LlmCall.Dialect.OpenAI.Responses.Connection.md): A WebSocket connection for OpenAI Responses (`wss://…/v1/responses`).
- [LlmCall.Dialect.OpenAI.Responses.Connection.Stream](LlmCall.Dialect.OpenAI.Responses.Connection.Stream.md): The lazy `Enumerable` over one Responses WebSocket lane.
- [LlmCall.EncodedRequest](LlmCall.EncodedRequest.md): A request rendered into one protocol's wire shape.
- [LlmCall.Error](LlmCall.Error.md): A codec error.
- [LlmCall.HTTP](LlmCall.HTTP.md): HTTP transport, backed by Finch.
- [LlmCall.HTTP.Request](LlmCall.HTTP.Request.md): One HTTP request, fully resolved by the client.
- [LlmCall.JSON](LlmCall.JSON.md): The JSON seam.
- [LlmCall.Media](LlmCall.Media.md): Namespace for the media adapters vendored with this library.
- [LlmCall.Media.Capabilities](LlmCall.Media.Capabilities.md): The media surface a codec declares.
- [LlmCall.Media.Codec](LlmCall.Media.Codec.md): The behaviour every media generation adapter implements.
- [LlmCall.Media.Conformance](LlmCall.Media.Conformance.md): Fixture-based assertions for media codec authors.
- [LlmCall.Media.Dialect.Adapter](LlmCall.Media.Dialect.Adapter.md): `use` for a media adapter.
- [LlmCall.Media.Dialect.Gemini.Predict](LlmCall.Media.Dialect.Gemini.Predict.md): Gemini Imagen image generation (`POST /v1beta/models/{model}:predict`).
- [LlmCall.Media.Dialect.MiniMax.Video](LlmCall.Media.Dialect.MiniMax.Video.md): MiniMax Hailuo video generation (`v1`).
- [LlmCall.Media.Dialect.OpenAI.Images](LlmCall.Media.Dialect.OpenAI.Images.md): OpenAI image generation (`POST /v1/images/generations`).
- [LlmCall.Media.Dialect.Seedance.Video](LlmCall.Media.Dialect.Seedance.Video.md): Seedance video generation (BytePlus ModelArk content generation tasks).
- [LlmCall.Media.Request](LlmCall.Media.Request.md): A provider-neutral image or video generation request.
- [LlmCall.Media.Response](LlmCall.Media.Response.md): The result of a media generation.
- [LlmCall.Message](LlmCall.Message.md): One user or assistant turn.
- [LlmCall.Operation](LlmCall.Operation.md): A long-running media generation job.
- [LlmCall.Provider](LlmCall.Provider.md): Declares a provider: where it lives, how it authenticates, and which codecs it
serves.
- [LlmCall.Reasoning](LlmCall.Reasoning.md): Canonical reasoning controls.
- [LlmCall.Reasoning.Capabilities](LlmCall.Reasoning.Capabilities.md): The reasoning surface a codec and model accept.

- [LlmCall.Request](LlmCall.Request.md): A provider-neutral completion request.
- [LlmCall.Response](LlmCall.Response.md): A provider-neutral completion response.
- [LlmCall.SSE](LlmCall.SSE.md): Server-sent event framing.
- [LlmCall.Stream](LlmCall.Stream.md): A lazy `Enumerable` over a streaming call.
- [LlmCall.StreamAssembler](LlmCall.StreamAssembler.md): Turns streamed tool-call fragments into complete `LlmCall.ToolCall` values.
- [LlmCall.StreamDelta.MessageDelta](LlmCall.StreamDelta.MessageDelta.md): A stop reason, usage, or both, arriving mid-stream.
- [LlmCall.StreamDelta.Raw](LlmCall.StreamDelta.Raw.md): A provider payload the codec did not map to a canonical delta.
- [LlmCall.StreamDelta.Text](LlmCall.StreamDelta.Text.md): Incremental text.
- [LlmCall.StreamDelta.Thinking](LlmCall.StreamDelta.Thinking.md): Incremental reasoning text, a reasoning signature, or both.
- [LlmCall.StreamDelta.ToolCall](LlmCall.StreamDelta.ToolCall.md): A tool-call fragment.
- [LlmCall.StreamEvent](LlmCall.StreamEvent.md): A normalized streaming event.
- [LlmCall.StreamState](LlmCall.StreamState.md): Opaque per-codec streaming state.
- [LlmCall.ToolCall](LlmCall.ToolCall.md): A call the model asked for.
- [LlmCall.ToolResult](LlmCall.ToolResult.md): The result of a tool call, ready to send back to the model.
- [LlmCall.ToolSpec](LlmCall.ToolSpec.md): A tool the model may call.
- [LlmCall.Usage](LlmCall.Usage.md): Token accounting as reported by the provider.
- [LlmCall.WS](LlmCall.WS.md): The WebSocket transport contract.
- [LlmCall.WS.Mint](LlmCall.WS.Mint.md): The default WebSocket transport, backed by `Mint.WebSocket`.

- Exceptions
  - [LlmCall.HTTP.Error](LlmCall.HTTP.Error.md): Raised when a streaming call fails after `LlmCall.stream/3` has returned.
  - [LlmCall.WS.Error](LlmCall.WS.Error.md): Raised by a WebSocket stream when the session or the connection fails.

