TachiBot/Docs
HomeGitHubnpm
IntroductionInstallationQuick StartExamplesTools OverviewPlannerPrompt TechniquesWorkflowsPingPongTool ProfilesTools ConfigThemesUsage StatsAPI KeysWhat's NewTerms & ConditionsCookie Policy
Products
TachiBot MCPdokorotachi-agent
Getting started
  • Introduction
  • Installation
  • Quick Start
  • Examples
Core features
  • Tools Overview
  • Planner
  • Prompt Techniques
  • Workflows
  • PingPong
Configuration
  • Tool Profiles
  • Tools Config
  • Themes
  • Usage Stats
  • API Keys
Resources
  • What's New
Legal
  • Terms & Conditions
  • Cookie Policy

Loading documentation...

← Previous
Tool Profiles
Next →
Themes
Edit this page on GitHub
Docs/Configuration

Tools Configuration

Control which tools are loaded to manage token usage. Schema size varies a lot by tool — the median is about 250 tokens, ranging from roughly 60 to 940. Enable only what you need.

Why Configure Tools?

TachiBot provides 67 tools, but you rarely need all of them at once. By enabling only the tools you need, you can control token usage and costs:

Minimum
Just think tool = ~100 tokens
Core set
think + focus + nextThought = ~1,350 tokens
Full suite
All 67 tools enabled = ~17,300 tokens

Configuration File

Create or edit tools.config.json in your project root:

tools.config.jsonjson
1{
2 "tools": {
3 // Core Tools (3)
4 "think": true,
5 "focus": true,
6 "nextThought": true,
7
8 // Perplexity Tools (2)
9 "perplexity_ask": true,
10 "perplexity_reason": false,
11
12 // Grok Tools (6)
13 "grok_reason": false,
14 "grok_code": false,
15 "grok_debug": false,
16 "grok_architect": false,
17 "grok_brainstorm": false,
18 "grok_search": true,
19
20 // OpenAI Tools (4)
21 "openai_reason": false,
22 "openai_brainstorm": true,
23 "openai_code_review": false,
24 "openai_explain": false,
25
26 // Gemini Tools (3)
27 "gemini_brainstorm": true,
28 "gemini_analyze_code": false,
29 "gemini_analyze_text": false,
30
31 // OpenRouter Tools (2-3) - All require OPENROUTER_API_KEY
32 "qwen_coder": false, // Qwen3-Coder for code tasks
33 "kimi_thinking": false, // Moonshot Kimi K3 for agentic reasoning
34 "qwen_competitive": false, // Competitive programming / LeetCode-style problems
35
36 // Workflow Tools (7)
37 "workflow": true,
38 "list_workflows": true,
39 "create_workflow": true,
40 "visualize_workflow": false,
41 "workflow_start": false,
42 "continue_workflow": false,
43 "workflow_status": false,
44
45 // Workflow Validators (2)
46 "validate_workflow": false,
47 "validate_workflow_file": false
48
49 // Note: pingpong is a bundled YAML workflow, run via the "workflow" tool
50 // (e.g. workflow --name pingpong-debate-3rounds), not its own boolean key
51 }
52}

Recommended Configurations

Minimal Setup (Research & Thinking)

Best for: Basic research and reasoning tasks•~2,500 tokens
minimal-config.jsonjson
1{
2 "tools": {
3 "think": true,
4 "focus": true,
5 "perplexity_ask": true
6 }
7}

Developer Setup (Code + Research)

Best for: Software development with code analysis•~4,500 tokens
developer-config.jsonjson
1{
2 "tools": {
3 "think": true,
4 "focus": true,
5 "nextThought": true,
6 "perplexity_ask": true,
7 "grok_code": true,
8 "grok_debug": true,
9 "gemini_analyze_code": true,
10 "qwen_coder": true,
11 "diff_review": true
12 }
13}

Research Setup (Deep Investigation)

Best for: Research, fact-checking, and analysis•~5,000 tokens
research-config.jsonjson
1{
2 "tools": {
3 "think": true,
4 "focus": true,
5 "perplexity_ask": true,
6 "perplexity_reason": true,
7 "grok_search": true,
8 "openai_search": true,
9 "gemini_analyze_text": true,
10 "jury": true
11 }
12}

Architecture Setup (System Design)

Best for: Software architecture and system design•~6,000 tokens
architecture-config.jsonjson
1{
2 "tools": {
3 "think": true,
4 "focus": true,
5 "nextThought": true,
6 "grok_reason": true,
7 "grok_architect": true,
8 "grok_brainstorm": true,
9 "openai_brainstorm": true,
10 "gemini_brainstorm": true,
11 "gemini_judge": true,
12 "workflow": true
13 }
14}

Full Workflow Setup (Everything)

Best for: Complex multi-step workflows using all capabilities•~13,000 tokens
full-config.jsonjson
1{
2 "tools": {
3 "think": true,
4 "focus": true,
5 "nextThought": true,
6 "perplexity_ask": true,
7 "perplexity_reason": true,
8 "grok_reason": true,
9 "grok_code": true,
10 "grok_debug": true,
11 "grok_architect": true,
12 "grok_brainstorm": true,
13 "grok_search": true,
14 "openai_reason": true,
15 "openai_brainstorm": true,
16 "gemini_brainstorm": true,
17 "gemini_analyze_code": true,
18 "gemini_analyze_text": true,
19 "gemini_judge": true,
20 "jury": true,
21 "qwen_coder": true,
22 "workflow": true,
23 "list_workflows": true,
24 "create_workflow": true
25 }
26}

Tool Categories

Core Tools

think
Structured reasoning scratchpad
focus
Multi-model orchestration
nextThought
Sequential thinking

Advanced Tools

jury
Multi-model jury panel with Gemini judge synthesis
plan_critique
Adversarial red-team of an existing plan (pre-mortem, ranked risks)
security_review
OWASP/CWE security audit with taint/data-flow analysis

Workflow Tools

workflow
Execute YAML/JSON workflows (e.g. --name pingpong-debate-3rounds for multi-model debates)
list_workflows
List available workflows
create_workflow
Create custom workflows

Provider-Specific Tools

Perplexity Tools

perplexity_ask
Web search with up-to-date information
perplexity_reason
Complex reasoning grounded in live web search

Grok Tools

grok_reason
First principles reasoning
grok_code
Code analysis and optimization
grok_debug
Debugging assistance
grok_architect
System architecture and design
grok_brainstorm
Creative brainstorming
grok_search
Live web search

OpenAI Tools

openai_reason
Mathematical and scientific reasoning with GPT-5.6
openai_brainstorm
Creative ideation using GPT-5.6

Gemini Tools

gemini_brainstorm
Collaborative brainstorming
gemini_analyze_code
Code quality and security analysis
gemini_analyze_text
Text analysis and sentiment

Qwen Tools

qwen_coder
Advanced code generation with Qwen3-Coder-Next (80B/3B MoE)
i
API keys
See API Keys Setup Guide for detailed instructions on obtaining and configuring these API keys.

Dynamic Reconfiguration

You can modify tools.config.json at any time. Restart TachiBot to apply changes:

terminalshell
1# Edit config
2vim tools.config.json
3
4# Restart TachiBot (if running as daemon)
5tachibot restart
6
7# Or just start fresh
8tachibot --config tools.config.json

Best Practices

Start minimal
Enable fewer tools, add more as needed
Match to task
Use developer config for code, research config for analysis
Monitor tokens
Check your context usage in Claude Desktop
No API key? Disable it
Tools requiring unavailable keys will fail
Core tools always on
Keep think / focus / nextThought enabled for best performance
✓
Pro Tip
The think tool alone provides 54% improvement in complex reasoning tasks (per Anthropic research). Always keep it enabled.

Next Steps

API Keys Setup →
Configure provider credentials
Explore Tools →
See what each tool does
Learn Workflows →
Build custom orchestrations