-
-
Notifications
You must be signed in to change notification settings - Fork 4
FAQ
Frequently asked questions about local-first Refact.
No. Refact runs locally in your IDE and uses the providers or local runtimes you configure.
No. The VS Code extension and JetBrains plugin start the local refact-lsp engine. Configure model providers or local runtimes in the Refact UI.
Refact supports Anthropic, OpenAI, OpenAI Responses, OpenAI Codex, OpenRouter, Ollama, LM Studio, vLLM, Groq, DeepSeek, Doubao, xAI, Gemini, Qwen, Kimi, Zhipu, MiniMax, GitHub Copilot, Custom, and Claude Code provider flows.
Yes. Use Ollama, LM Studio, vLLM, or a custom OpenAI-compatible endpoint. Start the runtime, make the model available there, add the runtime in Refact, and select the model in Default Models.
Refact does not sell model access. Billing, quotas, rate limits, and model access are handled by the provider or local runtime you configure.
The normal setup path sends requests only to configured providers, endpoints, and integrations. Local project data such as trajectories, tasks, knowledge, provider settings, and usage summaries are stored locally.
The IDE UI and Refact engine run locally. A workflow is fully local only when the configured model runtime and enabled tools are local. Hosted model providers and external integrations require network access.
The engine includes local capability metadata and combines it with configured provider data. This is why the model documentation describes provider families and capabilities instead of a fixed exhaustive model list.
Start with one strong chat or agent model, one faster chat model if available, and one completion-capable model if you want inline suggestions. Add an embedding model when you use semantic search or knowledge workflows.
Yes. Add, remove, or disable providers in Provider Setup, then update Default Models. Existing local trajectories and task data remain in your project/user Refact directories.
Refact on GitHub: https://github.com/JegernOUTT/refact
- Agent Modes
- Agent Tools
- Task Planner & Cards
- Worktrees
- Subagents
- Memory & Knowledge
- Hidden Roles & Plans
- Context Compression
- Scheduler & Cron
- Processes & PTY
- Buddy
- MCP
- Skills, Commands & Hooks
- Marketplace
- Chat System
- Providers
- Caps & Models
- Code Completion (FIM)
- AST
- VecDB
- Exec Runtime
- HTTP API
- Checkpoints & Git
- Voice