Everything you need to know about Pongogo, how it works, and how to get the most out of it.
General
Why is it called "Pongogo"?
The name combines "Pongo" (the genus name for orangutans, one of our
closest primate relatives known for their intelligence and problem-solving abilities)
with "GoGo" (representing energy, movement, and enthusiasm). Together
they create "Pongogo"—a clever, energetic helper that swings through your codebase
with purpose!
What is Pongogo?
Pongogo is an app + service that seamlessly integrates your IDE with project management
to give solo founders and small teams an on-demand agentic feature team. It creates warm, personable
AI teammates that feel genuinely collaborative—without the coldness, sycophancy, or hallucinations
that undermine trust in typical AI interactions. Through interactive workflows and self-improving
patterns, it gets better at understanding your preferences and working style over time.
Who is Pongogo for?
Pongogo is designed for solo founders, indie developers, and small teams
who want an on-demand feature team without hiring. It's opinionated by design with specific
views on how software should be built: roles-as-code for EM/PM guidance, quality-first workflows
with tests and docs by default, and trust-based principles that ensure complete understanding
of your codebase.
How It Works
What is MCP?
MCP stands for Model Context Protocol. Pongogo is built around emerging
standards like MCP servers, instruction files for agents, and the latest IDE functionality
in VS Code/Copilot, Claude Code, Warp, ChatGPT Codex, and Cursor. Everything lives in your
repo and PRs—no platform switch required.
What does Pongogo do in practice?
Pongogo provides key functionality including:
Intelligent task routing—assigns work to cloud agents, client agents, or humans based on complexity
Self-improving knowledge—learns from retros and creates guardrails to prevent repeating mistakes
Agentic PM—manages roadmaps, breaks work into chunks, knows what can run in parallel
Trust-based principles—ensures agents truly read and understand context, no shortcuts
Prototype sprints—from idea to MVP with clean, maintainable code and decision logs
How does the autonomy ladder work?
Pongogo uses a propose → patch → apply workflow with bounded autonomy.
Agents propose focused patches and scaffold tests alongside code. You review diffs with
approvals at each step. Everything has audit trails and easy rollback. This keeps you
in control while letting agents do the heavy lifting.
Installation & Setup
How do I install Pongogo?
The easiest way is via Homebrew on macOS:
$ brew install pongogo/tap/pongogo
You can also build from source. Check the documentation for detailed instructions.
What clients does Pongogo work with?
Pongogo works with any MCP-compatible client, including:
→ Claude Desktop
→ Cursor
→ Windsurf
→ Other MCP-compatible tools
How do I configure my AI client?
After installing Pongogo, start the MCP server with pongogo serve,
then add the server URL to your client's MCP configuration. Each client has slightly
different settings—check your client's documentation for details on adding MCP servers.
Customization
Can I add my own instruction files?
Yes! Pongogo is designed to be extensible. You can add custom instruction
files for your team's specific conventions, coding standards, or domain knowledge. These
will be indexed and routed alongside the built-in instructions.
Can I disable certain instruction categories?
Yes, you can configure which instruction categories are active. This is useful if you have
your own guidelines for certain areas or want to focus the agent on specific topics.
Still Have Questions?
Start with the install steps, then skim the sections above.
If you're still stuck, we'll help you get unstuck.