fa_llm 0.2.1 copy "fa_llm: ^0.2.1" to clipboard
fa_llm: ^0.2.1 copied to clipboard

Published 2 days ago
SDKDartFlutter
PlatformAndroidiOSLinuxmacOSwebWindows
→

Metadata

Reusable LLM provider adapters for Fa and related projects. Pure Dart core: OpenAI-compatible, OpenRouter, Ollama, and configuration.

More...

0.2.1 — 2026-09-30 #

  • LlmConfig.fromEnvironment base URLs (the explicit argument or {PROVIDER}_BASE_PATH/{PROVIDER}_BASE_URL) are normalized for the OpenAI-compatible providers: a bare origin or versioned base (…/v1) resolves to the full /v1/chat/completions endpoint, replacing the ollama-only fixup. Full endpoints, custom gateway paths, and query-string URLs pass through verbatim; copilot base URLs are never rewritten. The resolution order and accepted shapes are now documented in the class docs and the package README.

0.2.0 — 2026-08-28 #

  • GitHub Copilot as a first-class provider: OAuth device-flow auth (request/poll with pending/slow_down/expired/denied semantics), GitHub-token-to-Copilot-token exchange with dead-token detection, and login resolution.
  • CopilotTokenManager: cached short-lived Copilot tokens with proactive refresh (2 min before expiry), 30 s minimum exchange spacing, single-flight locking, and forced refresh (getAgain) for 401/403 retries.
  • CopilotTokenStore abstraction with an in-memory implementation for tests and apps without a secure backend.
  • CopilotProvider over /chat/completions: streaming SSE and complete responses, mandatory Copilot headers, per-request X-Initiator and Copilot-Vision-Request, cancel that stops emission immediately, refresh-once retry on 401/403, and listModels() parsing capabilities/limits/supported_endpoints.
  • ProviderFactory copilot branch with accountType (individual/business/enterprise), explicit baseUrl override, and entry-name-scoped token store wiring.
  • LlmConfig gains optional accountType and entryName fields.

0.1.0 #

  • Initial release.
  • OpenAI-compatible provider with streaming support.
  • OpenRouter provider with model routing.
  • Ollama provider for local inference.
  • Provider configuration and resolution utilities.
  • Token counting and context window management.

← Metadata

0
likes
140
points
29.2k
downloads

Documentation

API reference

Publisher

unverified uploader

Weekly Downloads

Metadata

Reusable LLM provider adapters for Fa and related projects. Pure Dart core: OpenAI-compatible, OpenRouter, Ollama, and configuration.

Homepage
Repository (GitHub)
View/report issues

License

MIT (license)

Dependencies

http, meta

More

Packages that depend on fa_llm

Back