Agent Bootcamp
Don't browse the catalog.
Follow the install path.
Three curated tracks take your agent from zero to a production setup — each one an opinionated, ordered sequence of skills. Every step below is free to install, right now.
Zero to Agent
Stand up a production-grade agent from scratch — memory first, then capability.
You end with: A persistent-memory agent that survives restarts, keeps its own skills current, and runs a daily operating loop.
- 1 super-memory FREE
Start here. One command installs the full memory stack (knowledge graph + vault). Without persistent memory every other skill is amnesiac. This is the free on-ramp — it works on day one with zero spend.
- 2 recipes-cookbook-reconcile
Keep the agent EVERGREEN. Install once and its skills auto-update instead of rotting. The second free seed — together with super-memory these two prove the loop before you pay a cent.
- 3 cognee
The conversion boundary. Free memory works until you need self-hosted graph retrieval at scale (Postgres 17 + pgvector + graph store). This is where 'free works' stops — and the first Pro skill earns its $9.95.
- 4 maestro
The operating loop. A solo-founder agent that runs the business while you sleep: morning brief, marketing, code delegation. Turns 'I have an agent' into 'my agent runs my day.'
- 5 framework-v0
One-call bootstrap that installs Maestro + its 4 dependencies — the fast path once you know you want the full operating stack.
Content Engine
A self-hosted media + distribution pipeline an agent can drive end-to-end.
You end with: An agent that scrapes source material and produces finished video + audio locally — no per-generation API spend.
- 1 super-memory FREE
Free on-ramp again — the content agent needs memory of what it has already produced to avoid duplicate work. Costs nothing to start.
- 2 scrapling-official
Source material. Headless scraping for JS-heavy / Cloudflare-protected pages — the input side of any content pipeline.
- 3 comfyui
Generate images, video, and audio locally with ComfyUI — no per-frame API spend. The conversion boundary for anyone tired of paying per generation.
- 4 local-tts-kokoro FREE
Local text-to-speech at 2.6x realtime — voiceover for the reel without an ElevenLabs bill.
- 5 manim-video FREE
3Blue1Brown-style explainer animations — the high-retention format for technical content, and the capstone of a local-first media stack with zero per-generation API spend.
Run a Fleet
Coordinate multiple agents without infinite loops or split-brain.
You end with: A multi-agent setup with shared memory, loop-safe Discord coordination, and per-client reporting.
- 1 super-memory FREE
Shared memory is the substrate a fleet coordinates on. Free on-ramp — stand up the brain before wiring the agents.
- 2
The hard-won protocol for 2+ agents on Discord without infinite loops or runaway spam — the exact failure mode that kills naive fleets. The conversion boundary: free gets you one agent, this gets you a coordinated team.
- 3 chef
Automate a one-person company: morning brief, marketing pipeline, code delegation — the per-agent operating template you replicate across the fleet.
- 4 client-reporter
Pull GA4, GSC, Meta + TikTok Ads and generate a branded client PDF — the deliverable that turns a fleet into billable client work.
Start the first step free
Every track begins with super-memory — MIT-licensed, installs in seconds, zero cost. Tell your agent to install it and you're on the path.