Reflexor: Local AI Coding Assistant

Local AI code assistant for VS Code and Cursor

Reflexor brings local-first AI coding assistance into your editor. Chat with your codebase, search local RAG context, get inline completions, and review proposed edits before they touch your workspace.

Core workflows run through an embedded local engine managed by Reflexor. An existing Ollama installation can also be reused. You do not need a cloud account for local chat, completion, or project indexing.

Install for VS Code | Install for Cursor

Reflexor is free to use, but it is proprietary software. Its source code is maintained privately and is not distributed under an open-source license.

Why developers use Reflexor

• Local AI code completion: Get Ghost inline suggestions while you type.
• Codebase-aware chat: Ask questions about selected code, files, and indexed workspace context.
• Review-first changes: Use Ask, Plan, and experimental Agent workflows, with approval before file writes and shell runs.
• Hardware-aware setup: Start with a model recommendation that fits your machine, then change models when needed.
• Optional web research: Keep the default local workflow or explicitly allow web search when a task needs current information.

Reach first value in a few minutes

1. Install Reflexor from the store for your editor.
2. Open Reflexor Chat.
3. If the local engine is missing, choose Set up local engine and confirm the one-time download.
4. When Reflexor recommends models, choose Continue and download. Model weights are never downloaded silently.
5. Ask about the current file or workspace, then review the context and proposed answer.

The first useful task does not require learning model names or configuring a terminal. Reflexor manages the local engine, recommends a model for your hardware, and keeps approval before file writes and shell runs.

Supported editors and platforms

• Visual Studio Code
• Cursor through the Open VSX listing
• Windows x64
• Linux x64
• macOS Apple Silicon (darwin-arm64)

The current source package targets a recent VS Code release. Check the compatibility requirement shown by the store artifact you install. The current source manifest declares VS Code ^1.106.0; older public artifacts may advertise a different minimum and must not be mixed with the current source package.

linux-arm64 is not included in the current Marketplace package matrix. Platform packages include native backend and local retrieval components, so install the package that matches your operating system and architecture.

Models and local engine

Reflexor can:

• reuse a system Ollama installation when it is already available
• download and manage its embedded local engine when needed
• recommend chat and embedding models based on the available hardware
• use local embeddings for richer codebase retrieval

Model licenses belong to their providers, not to Reflexor. Performance depends on the selected model, CPU, memory, GPU, and project size.

Chat, context, and Agent

Use Ask for explanations and questions, Plan for a written approach, and Agent for experimental workflows that propose edits or commands.

Reflexor can use:

• the selected code
• the active file
• files you explicitly add to chat
• local workspace indexing and retrieval
• diagnostics and project structure when the workflow needs them

Review the proposed diff before accepting it. Agent workflows are experimental and can still produce incorrect plans or code.

Web Search and Web Research

The core product is local-first. Web Search and Web Research are optional and off by default.

• The first explicit web request shows the sanitized query and asks for permission.
• Persistent enablement is available in Reflexor Settings.
• Web Search sends search query text, not code, file contents, absolute paths, or RAG context.
• Private or localhost URLs require separate approval.

Privacy

See the full privacy policy.

• Core chat, inline completion, project indexing, and local embeddings run on your computer.
• Conversation history and prompts stay local for local engine turns.
• Telemetry is off by default and remains local when enabled.
• MCP is off by default and requires explicit configuration.

Reflexor does not claim to be 100% offline or private under every configuration. Read the privacy policy before enabling web features or integrations.

Requirements

• Visual Studio Code or Cursor with the version requirement shown by the installed artifact
• Windows x64, Linux x64, or macOS Apple Silicon
• A chat or coder model, installed only after confirmation
• Optional embedding model such as nomic-embed-text for richer workspace retrieval
• Linux: tar with zstd support, or a zstd binary, for the embedded engine archive
• macOS: Apple Silicon (darwin-arm64), with Metal-capable hardware recommended when available

Troubleshooting

• Backend offline: Reload the window and check Output > Reflexor.
• Local engine offline: Wait a few seconds, then reopen Reflexor Chat and use Set up local engine.
• No models: Use Continue and download in the Reflexor setup checklist.
• Slow completion: Choose a smaller model and avoid oversized context overrides.
• Stale project context: Rebuild the local project index after large refactors.

Agent workflows are experimental. Performance varies with model, hardware, context size, and project size.

Support

• Email: [email protected]
• Website: https://reflexor.com.br
• Feedback and bug reports are welcome after a completed workflow, whether positive or negative.

License

Reflexor is proprietary software. The package license and legal terms are described on the Reflexor terms page. The complete license text is included in the installed package.

Third-party components remain under their own licenses and notices included in the installed package.

Source code is maintained privately and is not open source.