https://github.com/cognix-dev/cognix
Autonomous code generation powered by flow engineering.
https://github.com/cognix-dev/cognix
ai aider alphacodium assistant claude cli code-generation developer-tools development-tools flow-engineering gpt llm multi-model python
Last synced: 3 months ago
JSON representation
Autonomous code generation powered by flow engineering.
- Host: GitHub
- URL: https://github.com/cognix-dev/cognix
- Owner: cognix-dev
- License: apache-2.0
- Created: 2025-08-13T05:49:26.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2026-02-14T14:22:12.000Z (3 months ago)
- Last Synced: 2026-02-14T22:47:46.652Z (3 months ago)
- Topics: ai, aider, alphacodium, assistant, claude, cli, code-generation, developer-tools, development-tools, flow-engineering, gpt, llm, multi-model, python
- Language: Python
- Homepage: https://cognix-dev.github.io/cognix/
- Size: 1.27 MB
- Stars: 9
- Watchers: 1
- Forks: 0
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGELOG.md
- License: LICENSE
Awesome Lists containing this project
README
# Cognix
Autonomous code generation powered by flow engineering.
[](https://github.com/cognix-dev/cognix)
[](LICENSE)
[](https://python.org)
---
## Quick Start
### 1. Install and run
```bash
pipx install cognix
cognix
```
### 2. First-time setup
When you run Cognix for the first time, an interactive wizard will help you set up your API key:
- Choose your AI provider (Anthropic, OpenAI, or OpenRouter)
- Enter your API key
- The wizard creates a `~/.cognix/.env` file automatically
### 3. Generate code
Try the included sample first (use `@` to specify a file):
```bash
cognix> /make @sample_spec_tetris.md
```
Or describe what you want to build:
```bash
cognix> /make "landing page with HTML and CSS"
```
A sample specification file `sample_spec_tetris.md` is included in the repository. Use it as a reference for writing your own specifications.
### 4. Available commands
Type `/help` in the CLI to see all available commands.
---
## API Key Setup
### Automatic setup (recommended)
Just run `cognix` and follow the interactive wizard.
### Manual setup
Edit the `~/.cognix/.env` file (Windows: `C:\Users\\.cognix\.env`):
**Anthropic Claude (default):**
```bash
ANTHROPIC_API_KEY=sk-ant-your_key_here
```
Get your key at: https://console.anthropic.com/
Supported models: Sonnet 4.5 (default), Opus 4.6, Opus 4.5
**OpenAI:**
```bash
OPENAI_API_KEY=sk-your_key_here
```
Get your key at: https://platform.openai.com/api-keys
Supported models: GPT-5.2, GPT-5.2 Codex
**OpenRouter:**
```bash
OPENAI_API_KEY=sk-or-v1-your_key_here
OPENAI_BASE_URL=https://openrouter.ai/api/v1
```
Get your key at: https://openrouter.ai/keys
### Switch models
```bash
cognix> /model
```
---
## MCP Server Integration
Use Cognix from Claude Desktop, Cursor, VSCode, or any MCP-compatible tool.
Add to `claude_desktop_config.json`:
```json
{
"mcpServers": {
"cognix": {
"command": "cognix-mcp"
}
}
}
```
---
## Data Storage
Cognix stores data in `~/.cognix/`:
```
~/.cognix/
├── .env # API keys & credentials
├── config.json # Your settings
├── memory.json # Conversation & project memory
├── repository_data.json # Repository analysis cache
├── ui-knowledge.json # UI component knowledge
├── app_patterns.json # App pattern definitions
├── default_file_reference_rules.md # File reference rules
├── sessions/ # Saved work sessions
├── backups/ # Automatic backups
├── logs/ # Debug logs
├── temp/ # Temporary files
└── impact_analysis/ # Code impact analysis results
```
**Privacy:** No telemetry. API calls only go to your configured LLM provider.
---
## System Requirements
- **OS:** Windows 10+, macOS 10.15+, or Linux
- **Python:** 3.9 or higher
- **Internet:** Required for LLM API access
---
## Links
- **Documentation:** [github.com/cognix-dev/cognix](https://github.com/cognix-dev/cognix)
- **Report Issues:** [GitHub Issues](https://github.com/cognix-dev/cognix/issues)
- **Discussions:** [GitHub Discussions](https://github.com/cognix-dev/cognix/discussions)
---
## License
Apache-2.0 License - see [LICENSE](LICENSE) file for details