Bring your own LLM
Point the app at an OpenAI-compatible base URL with your API key and model ID. Supported examples include OpenAI, OpenRouter, Ollama, LM Studio, and custom gateways.
OpenPilot
OpenPilot is a local, open-source desktop AI agent that connects to the OpenAI-compatible model endpoint you choose. It helps developers work in a selected project folder with file operations, terminal commands, optional web research, MCP extensions, skills, and project memory.
Local desktop agent with your own model
OpenPilot is a local, open-source AI agent harness for desktop use. Rather than supplying a hosted model or requiring a vendor account, it connects to an OpenAI-compatible endpoint that you configure, including cloud providers, local servers, and custom gateways.
After you select a workspace, the agent can work with files and run shell commands in that folder. Its tool set also includes optional live web research through TinyFish, MCP server connections, reusable skills, and project preferences stored in AGENTS.md. The intended workflow is to describe an outcome, review the tool activity, and refine the result with follow-up instructions.
Core capabilities
Workspace file operations
Within a selected workspace, the agent can read, create, edit, grep, and glob files, making it suitable for scaffolding projects and revising existing code.
Live terminal access
Run installations, development servers, and tests from the chat while streaming terminal output back into the conversation.
Web research
With TinyFish configured, OpenPilot can search and fetch live web pages for documentation, API references, error messages, and other external information.
Skills and project memory
Add reusable playbooks as skills and store project preferences in AGENTS.md so the agent can load them into context and update memory when requested.
MCP extensions
Connect MCP servers through mcp.json and toggle them per chat to extend the tool set with systems such as databases, browsers, and internal APIs.
Practical use cases
Build a project from a brief
Give the agent a project folder and ask it to scaffold an application, create configuration and source files, then refine the implementation through follow-up prompts.
Debug and verify code
Use the workspace and shell tools to investigate a failing test, modify the relevant files, install dependencies, and rerun the test suite while reviewing terminal output.
Research external technical information
Ask for current library documentation, API references, or explanations of external errors when the answer is not contained in the local repository. TinyFish supplies the web search and fetch tools for this workflow.
Apply consistent project conventions
Save conventions such as language preferences, package-manager choices, and repository rules in AGENTS.md so they remain available across later tasks.
Extend a development workspace
Attach MCP servers for additional tools or resources, then enable them only in chats that need access to databases, browsers, documentation, or internal APIs.
Pros and Cons
Pros
- Works with cloud APIs, local OpenAI-compatible servers, and custom gateways instead of requiring an OpenPilot account.
- Combines filesystem operations, shell commands, optional web research, MCP, skills, and project memory in one chat-based workflow.
- Runs locally and stores API keys in operating-system secure storage according to the product FAQ.
- MIT licensed and available as both a Windows installer and portable executable.
Cons
- A compatible model endpoint and API key are required; OpenPilot does not provide a built-in model marketplace.
- Live web search and page fetching require a separately configured TinyFish API key.
- The currently documented desktop release is Windows-focused; macOS builds are listed as coming soon.
FAQ
Which platforms does OpenPilot support?
OpenPilot runs as a local Electron desktop app. Windows 10 or later on x64 is supported now through an installer or portable executable; macOS builds for Apple silicon and Intel are listed as coming soon.
How do I connect a model?
Open the model settings and enter an OpenAI-compatible base URL, API key, and model ID. The documented examples include OpenAI, OpenRouter, Ollama, LM Studio, and custom gateways.
Can OpenPilot use MCP servers?
Yes. Add an MCP server to mcp.json, restart OpenPilot, and enable the server for individual chats. Tools and resources are supported; MCP prompts are described as coming next.
Where are API keys stored?
The app stores keys in operating-system secure storage: Keychain on macOS and DPAPI on Windows. The product site says OpenPilot has no servers and does not send keys to OpenPilot.
Is a separate service required for web research?
OpenPilot can perform local filesystem and terminal work without TinyFish. A TinyFish API key is required before the web_search and web_fetch tools become available for live web research.
Quick Facts
- Category
- AI agent / developer tool
- Application
- Local Electron desktop app
- Current platform
- Windows 10 or later, x64
- Model requirement
- Any OpenAI-compatible endpoint
- License
- MIT
- Web research
- Available through TinyFish configuration
Alternativas a OpenPilot
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
GitBot
GitBot is a local browser-based workspace for creating reusable coding bots powered by Claude Code, Codex, or OpenCode. It helps developers run those agents across repository folders with persistent threads and configurable permissions.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
hob
hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.
Manta AI
Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.
SonOf
SonOf connects to your repo and PM tool, audits the codebase and surrounding product context, and turns approved work into shipped tickets with senior engineering review. It is aimed at founders and engineering leaders who need backlog help without hiring a full team immediately.