https://github.com/caiopizzol/conclave
Multi-model code review command for Claude Code. Run AI reviews locally before pushing.
https://github.com/caiopizzol/conclave
ai claude-code cli code-review codex developer-tools gemini multi-model
Last synced: about 5 hours ago
JSON representation
Multi-model code review command for Claude Code. Run AI reviews locally before pushing.
- Host: GitHub
- URL: https://github.com/caiopizzol/conclave
- Owner: caiopizzol
- License: mit
- Created: 2026-01-15T13:40:15.000Z (5 days ago)
- Default Branch: main
- Last Pushed: 2026-01-16T19:45:23.000Z (3 days ago)
- Last Synced: 2026-01-18T11:14:38.132Z (2 days ago)
- Topics: ai, claude-code, cli, code-review, codex, developer-tools, gemini, multi-model
- Language: HTML
- Homepage: https://caiopizzol.github.io/conclave/
- Size: 94.7 KB
- Stars: 5
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
[](https://github.com/caiopizzol/conclave/releases)
Multi-model code review for [Claude Code](https://claude.com/claude-code). Run reviews across multiple AI CLI tools in parallel and get consensus-driven feedback.
## How It Works
```
/review
│
├── Claude Opus ──► reviews independently
├── Codex ────────► reviews independently
├── Gemini ───────► reviews independently
├── Qwen Code ────► reviews independently
├── Mistral Vibe ─► reviews independently
├── Ollama ───────► reviews independently
└── Grok ─────────► reviews independently
▼
Synthesis: consensus highlighted, noise filtered
```
When multiple models flag the same issue, that's a stronger signal than any single review.
Enable/disable any combination of tools to get diverse perspectives from different training datasets.
## Inspiration
Inspired by [LLM Council](https://github.com/karpathy/llm-council) — the idea that multiple LLMs reviewing the same problem surfaces stronger signals than any single model.
## Installation
```bash
git clone https://github.com/caiopizzol/conclave ~/dev/conclave
cd ~/dev/conclave
bun run register
```
To unregister:
```bash
bun run unregister
```
## Configuration
### Tools (`~/.config/conclave/tools.json`)
```json
{
"tools": {
"codex": {
"enabled": true,
"command": "codex exec --full-auto -",
"model": "gpt-5.2-codex",
"description": "OpenAI Codex CLI"
},
"claude-opus": {
"enabled": true,
"command": "claude --print",
"model": "opus",
"description": "Claude Code (Opus)"
},
"claude-sonnet": {
"enabled": false,
"command": "claude --print",
"model": "sonnet",
"description": "Claude Code (Sonnet)"
},
"gemini": {
"enabled": false,
"command": "gemini -o text",
"description": "Google Gemini CLI (uses default model)"
},
"qwen": {
"enabled": false,
"command": "qwen -o text",
"description": "Qwen Code (Alibaba)"
},
"mistral": {
"enabled": false,
"command": "vibe --output text -p",
"description": "Mistral Vibe (Devstral)"
},
"grok": {
"enabled": false,
"command": "grok -p",
"model": "grok-code-fast-1",
"description": "xAI Grok CLI (community)"
},
"ollama-qwen": {
"enabled": false,
"command": "ollama run",
"model": "qwen3-coder:480b-cloud",
"description": "Ollama (Qwen3 Coder 480B)"
},
"ollama-devstral": {
"enabled": false,
"command": "ollama run",
"model": "devstral-2:123b-cloud",
"description": "Ollama (Devstral 2 123B)"
},
"ollama-local": {
"enabled": false,
"command": "ollama run",
"model": "qwen2.5-coder:7b",
"description": "Ollama (Qwen 2.5 Coder 7B, local)"
}
},
"prompt_file": "~/.config/conclave/prompt.md"
}
```
You can define multiple entries for the same provider with different models (e.g., `claude-opus` and `claude-sonnet`).
The `model` field is optional for most tools. If omitted, each tool uses its default model. **Exception:** Ollama requires an explicit `model` since it has no default.
**Supported models:**
| Tool | Models | Documentation |
| ------- | ------------------------------------------------------------------------------------------------------------------------------ | -------------------------------------------------------------- |
| Codex | `gpt-5.2-codex`, `gpt-5.1-codex-mini`, `gpt-5.1-codex-max`, `gpt-5.2` | [Codex Models](https://developers.openai.com/codex/models/) |
| Claude | `opus`, `sonnet`, `haiku` (aliases) or full names like `claude-opus-4-5-20251101` | [CLI Reference](https://code.claude.com/docs/en/cli-reference) |
| Gemini | `gemini-2.5-pro`, `gemini-2.5-flash`, `gemini-3-pro-preview`, `gemini-3-flash-preview` | [Gemini CLI](https://geminicli.com/docs/cli/model/) |
| Qwen | `coder-model` (default), `vision-model` | [Qwen Code Docs](https://qwenlm.github.io/qwen-code-docs/) |
| Mistral | Config-based (`~/.vibe/config.toml`) | [Mistral Vibe Docs](https://docs.mistral.ai/mistral-vibe/) |
| Grok | `grok-code-fast-1`, `grok-4-1-fast-*`, `grok-4-fast-*`, `grok-3`, `grok-3-mini` | [xAI API Models](https://docs.x.ai/docs/models) |
| Ollama | `qwen3-coder:480b-cloud`, `devstral-2:123b-cloud`, or any model from library | [Ollama Library](https://ollama.com/library) |
> **Note:** Ollama cloud models use `:cloud` suffix and require `OLLAMA_API_KEY` environment variable. Get your API key at [ollama.com](https://ollama.com). You can also run local models (e.g., `qwen2.5-coder:7b`), but they are slow and require significant memory (~8GB+ RAM for 7B models).
> **Note:** Mistral and Grok use command-line argument passing (not stdin), which has a ~200KB limit on macOS. Very large diffs may cause these tools to fail while other tools succeed.
> **Note:** Grok uses the community CLI ([`@vibe-kit/grok-cli`](https://github.com/superagent-ai/grok-cli)) until xAI releases the official "Grok Build" CLI.
### Prompt (`~/.config/conclave/prompt.md`)
Customize review instructions with template variables:
- `{{branch}}` — current branch
- `{{target_branch}}` — target branch
- `{{diff}}` — the diff content
### Authentication
| Tool | Install |
| ------- | ----------------------------------------------------------------------------- |
| Codex | `npm install -g @openai/codex` |
| Claude | Built-in |
| Gemini | `npm install -g @google/gemini-cli` |
| Qwen | `npm install -g @qwen-code/qwen-code` |
| Mistral | `pipx install mistral-vibe` |
| Grok | `bun add -g @vibe-kit/grok-cli`; `export GROK_API_KEY="key"` in `~/.zshrc` |
| Ollama | [ollama.com/download](https://ollama.com/download); cloud: `export OLLAMA_API_KEY="key"` in `~/.zshrc`; local: `ollama pull ` |
## Usage
```bash
/review
```
## Philosophy
More models ≠ better. The value is **consensus**:
- 1 model flags issue → might be noise
- 2+ models flag same issue → likely real
Conclave surfaces what matters.
## License
MIT