ManagePrompt
ManagePrompt is a platform designed to assist in the acceleration and simplification of AI projects. It supplies the necessary tools and ta...
Last verified:
What is ManagePrompt?
ManagePrompt is a local LLM call debugger that captures every LLM API call during development. It provides full request/response details, token usage, cost estimation, and latency metrics all in a local web UI powered by WebSockets for real-time updates. The tool is designed for developers working with LLMs who need to debug and monitor their AI calls without sending data to external servers.
Key features include full capture of request and response bodies with visual flow visualization, token tracking for input/output/cache read/write tokens, real-time cost estimation using models.dev pricing data, automatic capture of both streaming and non-streaming calls, project-scoped SQLite databases per project directory with zero configuration needed, and integration via AI SDK middleware or a simple capture() function. It works with any LLM API including OpenAI, Anthropic, Google, Mistral, and more.
ManagePrompt is ideal for developers building AI applications, prompt engineers testing different prompts, and teams debugging LLM integration issues during development. The tool is open source, released under GNU GPL license, and can be installed via Homebrew, Go, or npm. No OpenAI keys are shared with any external service as it runs entirely locally.
ManagePrompt pricing
Pricing model: Paid
ManagePrompt is open source software released under GNU GPL license with no paid tiers. The tool is free to use. Installation is available via Homebrew (brew install techulus/tap/manageprompt), Go (go install github.com/techulus/manage-prompt/cmd/manageprompt@latest), or npm (pnpm add manageprompt). There is no subscription cost as it runs entirely locally on your machine.
ManagePrompt pros
- Captures every LLM API call during development
- Full request and response body capture
- Visual flow visualization for calls
- Token tracking for input, output, cache read/write
- Real-time cost estimation via models.dev pricing
- Automatic streaming and non-streaming call capture
- WebSocket-powered real-time UI with live updates
- Project-scoped SQLite database per directory
- Zero configuration needed
- Two-line integration with AI SDK middleware
- Works with any LLM API (OpenAI, Anthropic, Google, Mistral)
- Open source under GNU GPL license
- Multiple installation options (Homebrew, Go, npm)
- Runs entirely locally - no external data sharing
- You never share OpenAI keys with ManagePrompt
ManagePrompt cons
- Only runs locally - no cloud collaboration features
- SQLite database may have scaling limitations
- No built-in prompt versioning or management
- Limited to development/debugging use cases
- No team sharing or multi-user access
- Requires Node.js >= 18.0.0 for some installations
- No explicit warranty or guarantee provided
- Developer bears no liability for damages
Frequently asked questions about ManagePrompt
What is ManagePrompt?
ManagePrompt is a local LLM call debugger that captures every LLM API call during development. It provides full request/response details, token usage, cost estimation, and latency metrics in a local web UI.
Is ManagePrompt open source?
Yes, ManagePrompt is an Open Source SaaS product released under the terms of the GNU General Public License (GPL).
How do I install ManagePrompt?
You can install the server via Homebrew (brew install techulus/tap/manageprompt) or Go (go install github.com/techulus/manage-prompt/cmd/manageprompt@latest). The npm package for capturing LLM calls is installed via pnpm add manageprompt.
Which LLM APIs does ManagePrompt work with?
ManagePrompt works with any LLM API including OpenAI, Anthropic, Google, Mistral, and more. It is model-agnostic.
How do I integrate ManagePrompt with my AI SDK?
Use the AI SDK Middleware integration: pnpm add manageprompt, then import { generateText, wrapLanguageModel } from 'ai' and import { devToolsMiddleware } from 'manageprompt'. Wrap your model with wrapLanguageModel and middleware: devToolsMiddleware().
Does ManagePrompt capture streaming calls?
Yes, ManagePrompt automatically captures both streaming and non-streaming calls without any additional configuration.
How is cost estimation calculated?
Real-time cost is calculated using models.dev pricing data, providing accurate cost estimation for each LLM call.
Where is my data stored?
Data is stored in a SQLite database per project directory. The tool runs entirely locally, so you never share your OpenAI keys with ManagePrompt.
Does ManagePrompt require configuration?
No, ManagePrompt requires zero configuration. It uses SQLite DB per project directory automatically.
What are the system requirements?
Node.js >= 18.0.0 is required for some installations. The server runs a local web UI accessible at localhost.