---

What's New in v2.14.7

$3

- gemini_judge — Science-backed LLM-as-a-Judge (arXiv:2411.15594). 4 modes: synthesize, evaluate, rank, resolve
- jury — Multi-model jury panel. Configurable jurors (grok, openai, qwen, kimi, perplexity, minimax) run in parallel, Gemini synthesizes the verdict. Based on "Replacing Judges with Juries" (Cohere, arXiv:2404.18796)

$3

- Fixed sonar-pro model ID (was accidentally using lightweight sonar)
- perplexity_research now uses sonar-deep-research — exhaustive multi-source reports in a single call

$3

qwen_coder now runs on Qwen3-Coder-Next (Feb 2026) — purpose-built for agentic coding:

| | Before (qwen3-coder) | After (qwen3-coder-next) |
|---|---|---|
| Params | 480B / ~35B active | 80B / 3B active |
| Context | 131K | 262K |
| SWE-Bench | 69.6% | >70% |
| Pricing | $0.22/$0.88 per M | $0.07/$0.30 per M |

3x cheaper, 2x context, better benchmarks. Falls back to legacy 480B on provider failure.

$3

| Tool | Capability | Highlight |
|------|-----------|-----------|
| kimi_thinking | Step-by-step reasoning | Agent Swarm architecture |
| kimi_code | Code generation & fixing | SWE-Bench 76.8% |
| kimi_decompose | Task decomposition | Dependency graphs, parallel subtasks |
| kimi_long_context | Document analysis | 256K context window |

$3

- minimax_code — SWE tasks at very low cost (72.5% SWE-Bench)
- minimax_agent — Agentic workflows (77.2% τ²-Bench)

$3

- qwen_reason — Heavy reasoning with Qwen3-Max-Thinking (>1T params, 98% HMMT math)

---

Key Features

$3

- 48 AI Tools across 7 providers — Perplexity, Grok, GPT-5, Gemini, Qwen, Kimi, MiniMax
- Multi-Model Council — planner_maker synthesizes plans from 5+ models
- Smart Routing — Automatic model selection for optimal results
- OpenRouter Gateway — Optional single API key for all providers

$3

- YAML-Based Workflows — Multi-step AI processes with dependency graphs
- Prompt Engineering — 14 research-backed techniques built-in
- Verification Checkpoints — 50% / 80% / 100% with automated quality scoring
- Parallel Execution — Run multiple models simultaneously

$3

| Profile | Tools | Best For |
|---------|-------|----------|
| Minimal | 12 | Quick tasks, low token budget |
| Research Power | 30 | Deep investigation, multi-source |
| Code Focus | 28 | Software development, SWE tasks |
| Balanced | 38 | General-purpose, mixed workflows |
| Heavy Coding (default) | 44 | Max code tools + agentic workflows |
| Full | 50 | Everything enabled |

$3

- Claude Code — First-class support
- Claude Desktop — Full integration
- Cursor — Works seamlessly
- TypeScript — Fully typed, extensible

---

Quick Start

$3

``bash
npm install -g tachibot-mcp
`

$3

Gateway Mode (Recommended) — 2 keys, all providers:

`json
{
"mcpServers": {
"tachibot": {
"command": "tachibot",
"env": {
"OPENROUTER_API_KEY": "sk-or-xxx",
"PERPLEXITY_API_KEY": "pplx-xxx",
"USE_OPENROUTER_GATEWAY": "true"
}
}
}
}
`

Direct Mode — One key per provider:

`json
{
"mcpServers": {
"tachibot": {
"command": "tachibot",
"env": {
"PERPLEXITY_API_KEY": "your-key",
"GROK_API_KEY": "your-key",
"OPENAI_API_KEY": "your-key",
"GOOGLE_API_KEY": "your-key",
"OPENROUTER_API_KEY": "your-key"
}
}
}
}
`

Get keys: OpenRouter | Perplexity

See Installation Guide for detailed instructions.

---

Tool Ecosystem (48 Tools)

$3


perplexity_ask · perplexity_research · perplexity_reason · grok_search · openai_search · gemini_search

$3

grok_reason · openai_reason · qwen_reason · kimi_thinking · kimi_decompose · planner_maker · planner_runner · list_plans

$3

kimi_code · grok_code · grok_debug · qwen_coder · qwen_algo · qwen_competitive · minimax_code · minimax_agent

$3

gemini_analyze_text · gemini_analyze_code · gemini_brainstorm · openai_brainstorm · openai_code_review · openai_explain · grok_brainstorm · grok_architect · kimi_long_context

$3

think · nextThought · focus · tachi · usage_stats

$3

workflow · workflow_start · continue_workflow · list_workflows · create_workflow · visualize_workflow · workflow_status · validate_workflow · validate_workflow_file

$3

list_prompt_techniques · preview_prompt_technique · execute_prompt_technique

$3

- Challenger — Critical analysis with multi-model fact-checking
- Verifier — Multi-model consensus verification
- Scout — Hybrid intelligence gathering

---

Example Usage

$3

`typescript
// Create a plan with multi-model council
planner_maker({ task: "Build a REST API with auth and tests", mode: "start" })
// → Grok searches → Qwen analyzes → Kimi decomposes → GPT critiques → Gemini synthesizes

// Execute with checkpoints
planner_runner({ plan: planContent, mode: "step", stepNum: 1 })
// → Automatic verification at 50%, 80% (kimi_decompose), and 100%
`

$3


`typescript
kimi_decompose({
task: "Migrate monolith to microservices",
depth: 3,
outputFormat: "dependencies"
})
// → Structured subtasks with IDs, parallel flags, acceptance criteria
`

$3

`typescript
kimi_code({
task: "review",
code: "function processPayment(amount, card) { ... }",
language: "typescript"
})
// → SWE-Bench 76.8% quality analysis
`

$3

`typescript
focus({
query: "Design a scalable event-driven architecture",
mode: "deep-reasoning",
models: ["grok", "gemini", "kimi"],
rounds: 5
})
``

---

Documentation

- Full Documentation
- Installation Guide
- Configuration
- Tools Reference
- Workflows Guide
- API Keys Guide
- Focus Modes

$3

- Claude Code Setup
- Claude Desktop Setup
- Both Platforms

---

Contributing

Contributions welcome! See CONTRIBUTING.md for guidelines.

- Report Issues
- Request Features

---