• Product
  • Latest
Sign InSign Up
HomeProductLatest
DocsCareersAbout
Sign InGet Started

Same-day shipping for your product team

© 2026 Wheel Go Fast, Inc. All Rights Reserved.

GitHubEmail support
Product
  • Studio
  • Taskmaster
Latest
  • Thoughts
  • Changes
Resources
  • Docs
  • Careers
  • About
Legal
  • Terms of Service
  • Privacy Policy
  • Cookie Policy
Taskmaster/Bring Your Own Key

Your AI. Your choice.

Use any AI provider — 15+ API keys supported, zero-key CLI tools, local models, or let Hamster handle everything. Taskmaster adapts to how you work.

AI Roles

Three-role AI system

Assign different models to different roles for optimal cost and performance. Main handles complex reasoning, Research queries live sources, and Fallback ensures reliability during rate limits.

Main Role

Primary model for task generation, updates, and analysis. Best for complex reasoning.

--set-main claude-sonnet-4
Research Role

Used with --research flag. Best for real-time web search and fresh information.

--set-research sonar-pro
Fallback Role

Automatic failover if main provider fails. Ensures reliability during rate limits.

--set-fallback gpt-4o-mini

Providers

Bring your own API key

Store keys in .env or configure via tm models --setup. Use the provider that works best for you — Anthropic, OpenAI, Google, Perplexity, and more.

AnthropicRec

Claude 4.5 Opus, Sonnet, Haiku

OpenAI

GPT-5, GPT-4.5, o3, o4-mini

Google

Gemini 3 Pro, 2.5 Pro/Flash

Perplexity

Sonar Pro, Sonar Reasoning

Groq

Llama 4, Kimi K2, Mixtral

xAI

Grok 3, Grok 4

OpenRouter

50+ models from all providers

Mistral

Mistral Large, Codestral

Zero Config

Use your existing subscriptions

Already paying for Claude Max, ChatGPT Plus, or Gemini Code Assist? Use those subscriptions directly — no separate API key needed.

Claude Code CLI
Claude Max/Pro
$0 extra

Use your existing Claude subscription via OAuth

task-master models --set-main sonnet --claude-code
Gemini CLI
Gemini Code Assist (Free tier available)
$0 extra

Leverage Google OAuth with free or paid tiers

task-master models --set-main gemini-3-pro --gemini-cli
Codex CLI
ChatGPT Plus/Pro/Business
$0 extra

Access GPT-5 through your ChatGPT subscription

task-master models --set-main gpt-5-codex --codex-cli
Grok CLI
X Premium/SuperGrok
$0 extra

Use Grok models via X Premium subscription

task-master models --set-main grok-4-latest --grok-cli

Privacy

Run completely local

Use Ollama or LM Studio to run models entirely on your machine. No API key, no internet required, no data leaves your device. Enterprise teams can use AWS Bedrock, Azure OpenAI, or Google Vertex AI.

Popular Local Models
qwen3:32bExcellent for code generation
llama3.3:latestGeneral purpose, fast
devstral:latestOptimized for development
codellama:34bCode-specialized
AWS Bedrock
IAM credentials
Azure OpenAI
Azure API key
Google Vertex AI
Service account

Key capabilities

  • 15+ API providers supported
  • Three-role model configuration (main, research, fallback)
  • Zero-key CLI tools with existing subscriptions
  • Local model support via Ollama and LM Studio
  • Enterprise infrastructure (Bedrock, Azure, Vertex)
  • Automatic failover between providers

Works with

MCP Server — Configure which provider your MCP server uses.

Research — Research role powers Perplexity queries.

Hamster Studio — Or let Hamster handle AI configuration entirely.

Configure your AI provider.

Run tm models --setup to get started.

Get started