Skip to content

Post-merge cleanup for OpenAICompatClient (#89) #95

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
antoinezambelli merged 1 commit into from
May 31, 2026
Merged

Post-merge cleanup for OpenAICompatClient (#89) #95

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 0 additions & 6 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,12 +2,6 @@

All notable changes to forge are documented here.

## [Unreleased]

### Added
- **OpenAI-compatible hosted-provider client (`OpenAICompatClient`).** Adapter for any backend exposing `/v1/chat/completions` with optional bearer auth — covers Cloudflare Workers AI, Fireworks, OpenRouter, OpenAI itself, and similar. Native function calling, SSE streaming, full `LLMClient` protocol surface. Supports `extra_headers` for provider quirks (e.g. OpenRouter's `HTTP-Referer` / `X-Title`) without a per-provider registry. Exported from `forge` and `forge.clients`. #88.
- **Hosted-providers section in [Backend Setup](docs/BACKEND_SETUP.md)** covering bearer-auth setup, the `get_context_length() → None` contract, and the per-model nature of function-calling support on hosted providers.

## [0.7.2] — 2026-05-24

vLLM backend support — serve AWQ/GPTQ and other vLLM-hosted models behind forge's guardrails, in both proxy modes and via `WorkflowRunner`.
Expand Down
4 changes: 3 additions & 1 deletion docs/BACKEND_SETUP.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,12 +1,14 @@
# Backend Setup

How to point forge at a backend. Forge supports four:
How to point forge at a backend. Forge supports six:

| Backend | Forge client | Native FC | Default port | Best for |
|---|---|---|---|---|
| llama-server | `LlamafileClient` | Yes (with `--jinja`) | 8080 | Recommended — top-10 eval configs |
| llamafile | `LlamafileClient` | No (prompt-injected fallback) | 8080 | Single binary, zero setup |
| Ollama | `OllamaClient` | Yes | 11434 | Easiest model management |
| vLLM | `VLLMClient` | Yes (server-side parser) | 8000 | AWQ/GPTQ, high-throughput serving |
| OpenAI-compatible | `OpenAICompatClient` | Per-model | (caller URL) | Hosted providers (Cloudflare, OpenRouter, …) |
| Anthropic | `AnthropicClient` | Yes | (API) | Frontier baseline |

Install instructions for each backend live with the upstream project. Below is what forge expects once a backend is running.
Expand Down
Loading