feat(providers): Integrate FreeBuff AI to the list

[noloverme]

Problem / Use Case

Currently, OmniRoute supports a specific set of AI providers. There is a platform called FreeBuff (https://freebuff.com/) that aggregates and provides access to various AI models and capabilities. Integrating FreeBuff would expand the available routing options for users who want to leverage this platform's specific availability, models, or cost benefits.

Proposed Solution

Integrate FreeBuff into OmniRoute as a supported AI provider.

1. Create a new provider module/adapter for FreeBuff.
2. Implement the standard interface methods required by OmniRoute (e.g., initialization, request routing, payload mapping).
3. Ensure base URL mapping pointing to https://freebuff.com/ (or their specific API endpoint if applicable).

Alternatives Considered

The alternative is using existing custom endpoint configurations if OmniRoute supports generic Open-AI compatible base URLs, but having native, first-class support for FreeBuff would simplify configuration and handling for end-users.

Acceptance Criteria

• A new provider configuration option for FreeBuff is available in OmniRoute.
• Requests can be successfully routed through the FreeBuff API.
• Error handling properly captures and bubbles up any platform-specific responses.
• Documentation is updated to include FreeBuff setup instructions.

Area

Provider Support

Related Provider(s)

FreeBuff (https://freebuff.com/)

Additional Context

Note: Exact usage limits, tier structures, and official API documentation constraints for FreeBuff are currently being evaluated, but the platform is actively available for testing and integration.

Expected Test Plan

• Unit tests for the FreeBuff provider adapter initialization and request/response payload parsing.
• Mock integration tests simulating API responses (success and error states) from https://freebuff.com/.
• End-to-end routing validation to confirm the provider integrates seamlessly with OmniRoute's core switching logic.

[diegosouzapw]

Hi @noloverme! Thanks for the suggestion — we'd be happy to look at adding FreeBuff AI. 🙏

To scope it we need the integration details:

1. The API base URL and whether it's OpenAI-compatible (Chat Completions shape).
2. The auth model (API key? cookie/session? anonymous?).
3. The model list and an example request/response (a curl works great).

If FreeBuff already exposes an OpenAI-compatible endpoint, you may be able to use it today via our generic custom-provider path (Dashboard → Providers → add a custom OpenAI-compatible provider with the base URL + key). If it needs a custom protocol, the details above let us build a dedicated adapter.

Looking forward to your response! 🚀

[chirag127]

Additional research / context for FreeBuff integration

Adding technical detail to help maintainers scope the work. Confirmed via direct page fetch + GitHub:

What FreeBuff actually is

FreeBuff is the free fork/relaunch of CodebuffAI (per schema.org markup on https://freebuff.com — sameAs: ["https://github.com/CodebuffAI/codebuff", "https://codebuff.com"]).

• Open source repo: https://github.com/CodebuffAI/codebuff
• Positioning: "free alternative to Claude Code, Cursor, OpenAI Codex, Lovable, Replit Agent, Bolt.new, Windsurf, Emergent, Devin"
• Product surface: Freebuff CLI (terminal coding agent) + Freebuff Web (full-stack app builder)
• Bundled subagents: 9 specialized (code review, browser testing, etc.)
• Pricing: $0 free tier confirmed in schema.org markup (price: "0", priceCurrency: "USD")
• Rating claim: 4.9/5 (512 reviews) per schema.org markup
• Platforms: macOS, Windows, Linux

Integration questions for maintainers / FreeBuff team

1. Does FreeBuff expose an OpenAI-compatible REST API? Or is it CLI-only?
   The homepage describes it as a CLI tool — not clear it has a public /v1/chat/completions surface. Check CodebuffAI/codebuff repo for server mode.

2. What models does FreeBuff route to underneath?
   As a "free Claude Code/Cursor alternative" it presumably has its own routing logic to upstream models. OmniRoute would need to know what model IDs to expose.

3. Auth method?
   Free tier suggests no API key — but if there's an API, how is rate limiting enforced? Per-IP? Session token? Sign-up required?

4. Is FreeBuff itself a router that internally uses providers OmniRoute already has?
   If FreeBuff is just routing to Pollinations / Cerebras / Groq under the hood, integrating it would create a circular route — OmniRoute → FreeBuff → providers OmniRoute already supports. Worth confirming before building the adapter.

Suggested next steps

1. Confirm API surface with CodebuffAI maintainers (open an issue at CodebuffAI/codebuff asking about programmatic API access if not already there)
2. If CLI-only: integration would need an ACP-style transport (OmniRoute already supports ACP per [Feature] Adding Alternative Agent Client Protocol (ACP) Support for CLI-Supported Providers #235), not a REST adapter
3. If API exists: standard provider adapter pattern applies (registry entry + auth + translator + tests)
4. If FreeBuff is itself a router: deprioritize — adding a router-of-routers creates circular dependency risk

Related

• [Feature] Add Provider FreeBuff #4251 (closed) — "Add Provider FreeBuff" — would be useful to know why this was closed (no API? legal concerns? duplicate of something?). Could a maintainer link the closing reason here so this issue doesn't repeat the same path?
• [Feature] Adding Alternative Agent Client Protocol (ACP) Support for CLI-Supported Providers #235 (closed, implemented) — ACP support; relevant if FreeBuff is CLI-only
• OpenCode (sst/opencode, 180K stars) is a similar OSS coding agent already detectable by OmniRoute via the "also ACP" badge — same pattern might apply to FreeBuff

[diegosouzapw]

Hi @noloverme (and thanks @chirag127 for the deep research)! 🙏

After investigating, FreeBuff isn't viable as a first-class provider:

• No public REST API. freebuff.com markets three end-user tools (a CLI, a web app-builder, a chat UI) — "No API key, no credit card. It just runs." There's no /v1/chat/completions / /v1/models developer surface, and the open-source upstream (CodebuffAI/codebuff) is a terminal agent that internally routes through OpenRouter, with FreeBuff's free tier substituting ad-funded access to DeepSeek / Kimi / MiniMax.
• ToS + circularity. FreeBuff's terms disallow automated/robot access, and structurally it's a consumer of the very models OmniRoute already supports directly — routing OmniRoute → FreeBuff → DeepSeek/Kimi/MiniMax would be circular.
• This is effectively a duplicate of [Feature] Add Provider FreeBuff #4251 (closed 2026-06-19 with the same conclusion); nothing has changed on freebuff.com or the upstream repo since.

Recommendation: use DeepSeek / Kimi / MiniMax directly in OmniRoute (all already supported) for the same underlying models.

Archiving rather than discarding — if FreeBuff ever ships a real public API, we'll revisit and tag you. Closing for now. 🙏

[chirag127]

Hi @noloverme (and thanks @chirag127 for the deep research)! 🙏

After investigating, FreeBuff isn't viable as a first-class provider:

• No public REST API. freebuff.com markets three end-user tools (a CLI, a web app-builder, a chat UI) — "No API key, no credit card. It just runs." There's no /v1/chat/completions / /v1/models developer surface, and the open-source upstream (CodebuffAI/codebuff) is a terminal agent that internally routes through OpenRouter, with FreeBuff's free tier substituting ad-funded access to DeepSeek / Kimi / MiniMax.
• ToS + circularity. FreeBuff's terms disallow automated/robot access, and structurally it's a consumer of the very models OmniRoute already supports directly — routing OmniRoute → FreeBuff → DeepSeek/Kimi/MiniMax would be circular.
• This is effectively a duplicate of [Feature] Add Provider FreeBuff #4251 (closed 2026-06-19 with the same conclusion); nothing has changed on freebuff.com or the upstream repo since.

Recommendation: use DeepSeek / Kimi / MiniMax directly in OmniRoute (all already supported) for the same underlying models.

Archiving rather than discarding — if FreeBuff ever ships a real public API, we'll revisit and tag you. Closing for now. 🙏

Can we use web cookies I don't think the web have the ads or maybe they are blocked by me

[jonlwheat2-gif]

Would this bring more context for freebuff? https://github.com/Quorinex/Freebuff2API

[chirag127]

Would this bring more context for freebuff? https://github.com/Quorinex/Freebuff2API

@diegosouzapw do you still disagree on adding freebuff

[shabeer]

@diegosouzapw - The primary reason for requesting freebuff is that the models are free to use - good models - atleast in limited mode. I do not have a subscription for DeepSeek / Kimi / MiniMax. Freebuff web might have APIs that will help in integration.

Regarding ads - we should pass-through the ads. No need to block ads.