Templates

Agent platforms and pre-built apps. Clone, deploy, and iterate with coding agents.

Clone a template, deploy to your cloud, and use coding agents like Claude Code to build and improve agents on top.

Agent Platforms

Infrastructure starters for building your own agents. Pick where you want to run, then add your agents on top.

Docker

Run locally or self-host anywhere Docker runs.

Railway

One-click cloud deployment. Best for getting started.

AWS

Full control with infrastructure-as-code.

Platform	Best for	You manage	Cost
Docker	Local dev, self-hosted production	Everything	Free
Railway	Fast cloud deployment, small teams	Just code	~$20/mo
AWS	Enterprise deployments, full control	Infrastructure-as-code	~$75/mo

Apps

Pre-built agent systems with self-learning, context retrieval, and production infrastructure already wired up. Clone the repo, configure your data sources, and you’re running.

Dash

Self-learning data agent. Connects to your databases, learns your schema, improves with every query.

Scout

Company intelligence agent. Navigates Slack, Drive, and wikis to answer questions.

Coda

Code companion. PR reviews, issue triage, architecture questions. Lives in Slack.

PAL

Personal agent that learns. Knowledge base, wiki, and structured data that compounds.

Gcode

Self-improving coding agent. Persistent workspace, git-based isolation, full audit trail.

What’s Included

Each app ships with capabilities that typically take weeks to build:

Capability	What It Does
Self-learning	Agent improves from feedback. Every conversation makes the next one better.
Context retrieval	RAG, knowledge graphs, multi-source navigation. Grounded answers, not hallucinations.
Production infrastructure	PostgreSQL, pgvector, webhook endpoints. Deploy to Docker, Railway, or AWS.
Interfaces	Slack, Telegram, WhatsApp ready. Add more with a few lines of config.

Improving Your Agents

Each app includes tools for iterating on agent quality:

App	Improvement Tools
Dash	Automated improvement loop (python -m evals improve)
Scout	Probe library with /loop support
Coda	Eval suite for routing and synthesis
PAL	Smoke tests and behavioral evals
Gcode	LearningMachine for persistent context