loading…
Search for a command to run...
loading…
Enables creating projects from uploaded files directly from AI assistants. Provides tools for obtaining upload URLs and creating project import jobs from previo
Enables creating projects from uploaded files directly from AI assistants. Provides tools for obtaining upload URLs and creating project import jobs from previously uploaded files.
The rad quote MCP server lets you create projects from uploaded files directly from your AI assistant.
Pick the install path that matches you:
| Your situation | Path |
|---|---|
| I use Claude Desktop and do not want to install Node or open a terminal | Claude Desktop UI — paste https://api.limits.run/mcp?key=<YOUR_API_KEY> into the Custom Connector form |
| I am a developer, I have Node 18+, I use Cursor / Codex / Windsurf / Claude Code | One-command CLI |
| I want a double-click install for Claude Desktop and have Node 18+ on my machine | .mcpb bundle |
npx @radquote/mcp-setup@alpha install --env=stage
Requires Node 18+. Detects every supported client on your machine and writes the configuration for it. By default it configures all of them; pass --client=<id> to target one, or --pick to choose interactively.
Supported clients: claude-code, claude-desktop, cursor, windsurf, codex.
More detail: packages/setup/README.md.
If you prefer to edit configuration files yourself, copy the snippet for your client from examples/. Replace <MCP_URL> with https://api.limits.run/mcp (stage) and <YOUR_API_KEY> with your rad quote workspace API key.
Three tools appear in your AI assistant:
get_upload_url — obtain a short-lived signed URL for uploading one file.create_project_import_job — create a new project by importing tasks from previously uploaded files; optionally copy rates from an existing project as a read-only template.get_project_import_job — look up the current status of an import job by id. When the job is complete, the response contains project_id of the newly created project; when it has failed, the response carries a human-readable error.And one orchestration prompt:
estimate_project_workflow — a single entry point that walks the agent through the full flow: upload the file, create the import job, poll the job every 15 seconds, and reply with a link to the resulting project (or with the failure reason). Arguments: file (required absolute local path), title, client_title, minimum_budget, source_project_id (all optional).The estimate_project_workflow prompt packages the full recipe so you do not have to spell out the steps to the agent. Invocation depends on the client:
/mcp__radquote-stage__estimate_project_workflow, then fill in file and the optional fields.estimate_project_workflow from your client's prompts menu and fill the arguments.The workflow requires the agent to PUT the file to a signed URL. That means it works in clients that can run shell commands (Claude Code, Cursor, Codex, Windsurf) but not in Claude Desktop's built-in chat, which cannot perform arbitrary HTTP uploads. In Claude Desktop, upload the file some other way and then invoke
create_project_import_jobwith its public URL directly.
For reference, the PUT step looks like:
curl -X PUT --upload-file ./my-file.xlsx \
-H "Content-Type: application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" \
"<upload_url>"
The response is HTTP 201 with a JSON body containing url — feed that into create_project_import_job.file.
Pre-release. The production endpoint URL and the MCP Registry listing are being finalized. Currently only the stage environment (https://api.limits.run/mcp) is wired up — intended for internal QA and partner integration. Running the CLI with --env=prod prints a clear "not configured" message until production ships.
See docs/stage.md for details on the stage environment.
Добавь это в claude_desktop_config.json и перезапусти Claude Desktop.
{
"mcpServers": {
"rad-quote-mcp": {
"command": "npx",
"args": []
}
}
}Web content fetching and conversion for efficient LLM usage.
Retrieval from AWS Knowledge Base using Bedrock Agent Runtime.
Provides auto-configuration for setting up an MCP server in Spring Boot applications.
A very streamlined mcp client that supports calling and monitoring stdio/sse/streamableHttp, and can also view request responses through the /logs page. It also