Open Source MCP Server

Stop Switching Tabs. Command Every LLM From One Prompt.

Seven models, seven training sets, seven perspectives. Perplexity searches, Grok reasons, Gemini critiques, Kimi thinks step-by-step — different minds, one answer.

Gateway:One OpenRouter key

BYOB:Your own provider keys

Perplexity: Always needs own key

Get Started →Read Docs

"Have Grok check Twitter for that error message""Ask Perplexity what changed in React 19 this week""Get Gemini to brainstorm, then have Kimi K2.5 and GPT-5.2 both analyze it"

Providers

35+

Tools

Prompt Techniques

Stop Trusting. Start Verifying.

Same question. Different approach. Better answer.

"What breaking changes are in React 19?"

One Model

Generic advice, might miss recent updates

No sources or documentation links

Could confuse React 18 vs 19 features

No way to verify accuracy

Unverified, possibly outdated

TachiBot

1. Run in Parallel

OpenAIGooglePerplexityOpenRouter

2. Cross-Verify

Models challenge each other, 3-200 rounds

3. Fact-Check

PerplexityGrok

Live sources with recency filters

Verified, with official sources

Get Started →

Building Blocks for Thinking

Mix and Match Your Reasoning Pipeline

Mix and match tools, prompt techniques, and workflows to build your reasoning pipeline.

New

Multi-Model Planner

Create bulletproof implementation plans. A council of 6 AI models — Grok searches for ground truth, Qwen analyzes code, Kimi K2.5 reasons step-by-step and decomposes into dependency-ordered subtasks, GPT-5.2 critiques gaps, Qwen 235B drafts the synthesis, and Gemini judges the final plan with quality scores and verification checkpoints.

planner_maker + planner_runner

Multi-Model Council

Run your question through 4-6 AI models simultaneously. Each analyzes from its unique perspective. A final judge synthesizes the best insights, scores confidence, and resolves conflicts.

council + judge

7 AI Providers

GPT-5.2, Gemini, Grok, Perplexity, Kimi K2.5, Qwen, and MiniMax — each with unique strengths. Gateway mode with one OpenRouter key, or BYOB with your own provider keys.

98%Qwen 235B·HMMT76.8%Kimi K2.5·SWE88.4%GPT-5.2·GPQA95%Gemini 3·AIME72.5%MiniMax·SWE

gateway + BYOB

New

22 Prompt Techniques

Research-backed patterns like first_principles, tree_of_thoughts, and council_of_experts. Preview the enhanced prompt before executing. Apply any technique to any model.

preview → execute

YAML Workflows

Chain unlimited steps. Variable interpolation, step dependencies, auto-distillation, comparison tables, and optional AI judge. PingPong debates, research pipelines, code reviews.

workflow + planner

Decision-Making Framework

How Seven Minds Think Together

A structured reasoning pipeline — ground in data, decompose, explore alternatives, stress test for holes, then judge.

01Ground Truth

Search real-time data from 4 providers. No thinking starts without facts.

PerplexityGrok SearchGemini SearchOpenAI Search

02Break Down

Decompose into atomic parts. Map dependencies, constraints, and execution order.

Kimi K2.5Qwen 235BQwen Algo

03Explore Paths

Generate alternative approaches from different training data and perspectives.

GPT-5.2GeminiGrokMiniMax

04Stress Test

Attack assumptions. Find holes, blind spots, and failure modes in every path.

GPT-5.2 CriticQwen Reason

05Judge

Synthesize the best elements from every model. Resolve conflicts. Score everything. Not 10? Here's why — and how to fix it.

Gemini 3 ProKimi K2.5

Confidence91%

Code Quality8.4/10

-1.6 — Cyclomatic complexity >10 in handler

Fix: Extract validation to utility class

Security7/10

-3 — No rate limiting on auth endpoints

Fix: Add express-rate-limit, 5 req/min on /login

Performance9/10

-1 — O(n²) nested loop on unindexed array

Fix: Use HashMap for O(1) lookup

Every deduction comes with a reason and a fix. No mystery scores.

Why these models

Each model was chosen for a specific strength. Different training data, different benchmarks, different blind spots.

98%

Qwen 235B

HMMT — Harvard-MIT math tournament. Proof-based, multi-step olympiad problems that PhD students struggle with.

95%

Gemini 3 Pro

AIME — American Invitational Math Exam. Top math competition, one level below the International Math Olympiad.

88.4%

GPT-5.2

GPQA Diamond — PhD-level science questions written by domain experts. Tests deep reasoning, not memorization.

76.8%

Kimi K2.5

SWE-Bench Verified — resolves real GitHub issues from open-source repos. Tops Gemini 3 Pro (76.2%). Open-weights.

72.5%

MiniMax M2.1

SWE-Bench — agentic code fixes. Best value at $0.27/M input tokens. 10x cheaper than GPT-5.2 for routine tasks.

1261

QwQ-32B

CodeElo — competitive programming rating. Solves Codeforces-style algorithm problems. Efficient at 32B parameters.

Real Workflow Output

5 Models. One Answer.

Real GPT-5 → GPT-5.2 migration analysis, ~3 minutes

Query

"Should I migrate from GPT-5 to GPT-5.2? Differences, breaking changes, migration steps."

Version DiscoveryPerplexity

GPT-5.2 released Nov 12 — automatic migration, backward compatible

Feature ComparisonGrok

2 modes, 8 personalities, 25% better coding, 15% better factuality

Migration ImpactGPT-5.2

No breaking changes — backward compatible until Q1 2026

Performance AnalysisGemini

30% latency reduction, 10% fewer hallucinations

Final RecommendationKimi K2.5

Low-risk upgrade — enable in staging, test 1 week, then production

Verdict: Safe to upgrade

No breaking changes, backward compatible, easy rollback

92%

Try TachiBot Yourself

Customize Everything

Toggle tools, write workflows, control exactly what loads

Profile System

Toggle tools on/off per project

tools.config.json

{
  "customProfile": {
    "tools": {
      "perplexity_ask": true,
      "grok_reason": true,
      "qwen_coder": false
    }
  }
}

1 tool~400 tokens

All 35+ toolsloaded dynamically

YAML Workflows

Chain models into multi-step pipelines

general-council.yaml

steps:
  - tool: grok_reason
    output: reasoning
  - tool: perplexity_ask
    output: research
  - tool: gemini_analyze_text
    output: patterns
  - tool: openai_brainstorm
    output: final

Pass outputs between steps, run in parallel, add variables. Save as .yaml or .json.

Profile docs →Workflow examples →

Works With Leading AI Providers

Always using the latest models from each provider

OpenAI

GPT-5.2 series

Google

Gemini 3 Pro

Perplexity

Sonar Pro search

xAI

Grok 4.1

OpenRouter

Qwen3-235B, Kimi K2.5

MiniMax

M2.1 agentic model

Anthropic

Claude models

Works best with Claude Code MCP integration

Get Started in Minutes

Add to Claude Desktop or any MCP client

1Install via npm

npm install -g tachibot-mcp

2Add to Claude Desktop config

{
  "mcpServers": {
    "tachibot": {
      "command": "tachibot-mcp",
      "args": ["--config", "~/.tachibot/config.json"]
    }
  }
}

~/.config/Claude/claude_desktop_config.json

3Start using

tachibot workflow run general-council "Your query here"

Documentation GitHub

Built in public. Backed by stars.

TachiBot is open source and actively maintained. If it helps your workflow, a star helps us keep going.

Star on GitHub Report a bug or request a feature