Skip to content

## Pluggable AI provider — Claude, GitHub Copilot, or OpenAI Codex

Latest
Latest

Choose a tag to compare

View all tags
@mickyarun mickyarun released this 02 Jul 09:42
v1.3.4
052c593
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Pluggable AI provider — Claude, GitHub Copilot, or OpenAI Codex

Each organization now chooses which agent CLI runs its codebase-aware work. Pick your provider in the setup wizard and Settings → AI Configuration; every agent task — specs, chat, design, estimation, scans, triage — runs on it.

Highlights

  • Three providers, one selector. Claude Code, GitHub Copilot, or OpenAI Codex, switchable per org. All reach the same BUD/code-graph tools through the built-in MCP bridge.
  • Flexible auth per provider. Hybrid mode inherits your host login (claude login, gh/Copilot, or codex login) so agents run on your existing subscription — or paste an API key/token, stored encrypted (Fernet AES-128).
  • In-wizard connection test for the selected provider before you commit.
  • Provider-native repo setup. The first scan writes the right config for your provider — Claude → .claude/, Codex → .codex/hooks.json + config.toml, Copilot → .github/hooks/, plus AGENTS.md — all wired to the same dev-activity / BUD-todo hooks so tracking works everywhere.
  • Live progress for Codex & Copilot — runs now stream tool activity instead of showing a static spinner.
  • Provider-aware Agent Prompts. Model/effort dropdowns reflect the active provider (Copilot runs on auto-selected models).

Notes

  • Claude behavior is unchanged; existing orgs keep running exactly as before.
  • Ollama (local models) is next on the roadmap.
Assets 4
Loading