Tachibot Mcp
Wire one MCP server so Claude Code can route tasks across Perplexity, Grok, OpenAI, and optional Gemini with YAML workflows instead of juggling separate clients.
Overview
Tachibot MCP is a MCP server for the Build phase that orchestrates Perplexity, Grok, OpenAI, and optional Gemini through 31 tools, YAML workflows, and five token profiles inside your coding agent.
What is this MCP server?
- 31 MCP tools for search, reasoning, code analysis, and comparison across providers
- YAML-defined workflows for repeatable multi-step agent routines
- 5 token-optimized profiles to trim context on heavy orchestration runs
- stdio npm package tachibot-mcp v2.0.5 for local Claude/Cursor registration
- Required keys: Perplexity, Grok/xAI, and OpenAI; optional Google Gemini for brainstorming
- 31 MCP tools documented in registry metadata
- Package version 2.0.5 via npm identifier tachibot-mcp
Community signal: 15 GitHub stars.
What problem does it solve?
Solo builders waste time context-switching between LLM APIs and hand-rolling prompts when one agent session needs search, code analysis, and cross-model comparison.
Who is it for?
Indie devs shipping with Claude Code who already have Perplexity, xAI, and OpenAI keys and want centralized multi-model tooling.
Skip if: Builders who only use one model vendor and do not want to manage three required API secrets on localhost.
What do I get? / Deliverables
After install, your agent calls a single MCP surface to run multi-model workflows with profile-based token control instead of ad hoc API juggling.
- 31 callable MCP tools wired into your agent session
- Reusable YAML workflow definitions for multi-step model pipelines
- Selectable token-optimized profiles for long orchestration runs
Recommended MCP Servers
Journey fit
Canonical shelf is Build because the value is extending your coding agent with orchestrated multi-model tools during implementation and debugging. agent-tooling fits MCP servers that add callable tools and profiles directly into the agent loop, not a single REST integration.
How it compares
MCP multi-model orchestration layer, not a single-purpose skill markdown workflow.
Common Questions / FAQ
Who is tachibot-mcp for?
Solo and small-team builders using Claude Code or Cursor who need Perplexity search, Grok debugging, and OpenAI analysis from one registered MCP server.
When should I use tachibot-mcp?
Use it during Build when you are extending agent-tooling for research, architecture review, or YAML-driven multi-step tasks across several LLM providers.
How do I add tachibot-mcp to my agent?
Install npm package tachibot-mcp, set PERPLEXITY_API_KEY, GROK_API_KEY, and OPENAI_API_KEY (optional GOOGLE_API_KEY), then add the stdio server entry in your host's MCP config.