Guide to configuring Jan for optimal performance and use cases.
Jan stores configuration and data in:
| Platform | Location |
|---|---|
| Windows | C:\Users\<Username>\AppData\Roaming\Jan |
| macOS | ~/Library/Application Support/Jan |
| Linux | ~/.config/Jan |
| Platform | Data Location |
|---|---|
| Windows | C:\Users\<Username>\AppData\Roaming\Jan\data |
| macOS | ~/Library/Application Support/Jan/data |
| Linux | ~/.config/Jan/data |
Linux:
export XDG_CONFIG_HOME=/home/username/custom_config
Settings → Local Engine → Llama.cpp:
| Setting | Description | Recommended |
|---|---|---|
| Backend | Inference backend | CUDA (NVIDIA), Metal (Apple), CPU |
| GPU Offload | Layers to offload to GPU | Max for best performance |
| Context Length | Model context window | 4096-8192 (model dependent) |
| Threads | CPU threads for inference | Match CPU cores |
Settings → Hardware → GPUs:
| GPU Type | Setting | Value |
|---|---|---|
| NVIDIA | CUDA | Enable for NVIDIA GPUs |
| AMD | ROCm | Enable for AMD GPUs |
| Intel | DirectML | Enable for Intel Arc |
| Apple | Metal | Enabled by default on M1/M2/M3 |
Settings → Local Engine → MLX:
| Setting | Description | Recommended |
|---|---|---|
| MLX Backend | Apple Silicon acceleration | Enabled by default |
| Memory Limit | Max memory for model | 80% of available RAM |
Hub Tab:
| Action | Description |
|---|---|
| Search Models | Find models from HuggingFace |
| Download | Download model files |
| Delete | Remove downloaded models |
| Import | Import custom GGUF models |
Per-Model Configuration:
| Parameter | Description | Typical Value |
|---|---|---|
| Temperature | Response randomness | 0.7 |
| Top P | Nucleus sampling | 0.9 |
| Top K | Top-K sampling | 40 |
| Max Tokens | Maximum response length | 2048 |
| Repeat Penalty | Prevent repetition | 1.1 |
| Stop Words | Custom stop sequences | Model-specific |
Settings → Local Engine → Llama.cpp → Auto-Fit:
Settings → API Server:
| Setting | Default | Description |
|---|---|---|
| Port | 1337 | API server listening port |
| Host | localhost | Binding address |
| CORS | Enabled | Cross-origin requests |
| Authentication | Disabled | API key requirement |
Jan provides these OpenAI-compatible endpoints:
| Endpoint | Method | Purpose |
|---|---|---|
/v1/chat/completions |
POST | Chat completions |
/v1/completions |
POST | Text completions |
/v1/models |
GET | List available models |
/v1/embeddings |
POST | Generate embeddings |
curl http://localhost:1337/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "jan-v1",
"messages": [
{"role": "system", "content": "You are helpful"},
{"role": "user", "content": "Hello!"}
],
"temperature": 0.7,
"max_tokens": 1024
}'
Enable API key authentication:
Settings → API Server → Authentication:
Authorization: Bearer YOUR_KEYcurl http://localhost:1337/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your-api-key" \
-d '{"model": "jan-v1", "messages": [...]}'
Settings → Remote APIs:
| Provider | Configuration |
|---|---|
| OpenAI | Add API key from platform.openai.com |
| Anthropic | Add API key from console.anthropic.com |
| Mistral | Add API key from console.mistral.ai |
| Groq | Add API key from console.groq.com |
| Google Gemini | Add API key from makersuite.google.com |
Create Assistant:
| Setting | Description |
|---|---|
| Name | Assistant display name |
| System Prompt | Base instructions for assistant |
| Model | Default model for assistant |
| Tools | Enabled tools (MCP, function calling) |
| Knowledge | Attached documents/files |
Settings → MCP:
| Setting | Description |
|---|---|
| Enable MCP | Toggle MCP integration |
| MCP Servers | Configure MCP server connections |
| Tool Permissions | Control tool access |
Settings → Threads:
| Setting | Description | Default |
|---|---|---|
| Max Threads | Maximum saved threads | 100 |
| Auto-Save | Auto-save conversations | Enabled |
| Export Format | Default export format | Markdown |
Settings → Threads → Delete All:
Settings → Display → Theme:
Settings → Display → Language:
Settings → Advanced:
| Setting | Description |
|---|---|
| Debug Logging | Enable verbose logs |
| Telemetry | Usage statistics (optional) |
| Auto-Update | Automatic updates |
| Beta Features | Enable experimental features |
Settings → Network → Proxy:
| Setting | Description |
|---|---|
| HTTP Proxy | HTTP proxy URL |
| HTTPS Proxy | HTTPS proxy URL |
| No Proxy | Bypass proxy for hosts |
Settings → Advanced → Export:
Settings → Advanced → Import:
Windows:
xcopy /E /I "C:\Users\%USERNAME%\AppData\Roaming\Jan" "D:\Backup\Jan"
macOS/Linux:
tar -czf jan-backup.tar.gz ~/Library/Application\ Support/Jan
# or
tar -czf jan-backup.tar.gz ~/.config/Jan
Any questions?
Feel free to contact us. Find all contact information on our contact page.