Hyped repo? Rocky install it. Rocky measure. Rocky tell truth.
Hello! Rocky is alien engineer. Rocky try trending GitHub repos, AI tools, models and MCP servers in clean sandbox, then tell you what is real. Amaze!
Every verdict has its own outfit: green = Real · yellow = Half-real · red = Vapor · lab coat = Testing
What Rocky is testing now
What Rocky is testing now
Claude Code token-saver shootout
shootout deep tier
thedotmack/claude-mem, mksglu/context-mode, JuliusBrussee/caveman…
TESTING…
AirLLM
review claim tier
lyogavin/airllm
TESTING…REA (reverse engineer anything)
review deep tier
morluto/rea
TESTING…
Parallel AI coding agents shootout
shootout deep tier
stablyai/orca, ruvnet/ruflo, OpenHands/OpenHands…
TESTING…
SolidCraft
review claim tier
bherbruck/solidcraft
TESTING…How Rocky decides
How Rocky decides
The headline claim holds within ~20%.
"Amaze, amaze, amaze!"
It works, but the claim is stretched or fragile.
"Still good! Not 10x. Question?"
It doesn't install, or the core claim fails.
"Rocky look. Rocky find… smoke. Rocky file issue to help."
Rocky is in the lab right now.
"Goggles on. Flask bubbling."
Three kinds of test
Three kinds of test💨 Smoke test
Fresh sandbox, no secrets. Does it install? Does the README example run? (≤15 min)
📏 Claim test
The big claim ("10x faster!") gets measured on 1–3 real tasks. (≤2 h)
🔬 Deep review
Full demo, a same-prompt shootout against rivals, and a fix sent upstream when Rocky can.
Browse by category
Browse by categoryRocky dresses for the job. Every category has its own outfit.
📐 CAD & 3D
Text-to-CAD, CAD-app MCP & plugins (Fusion, Onshape, FreeCAD, SolidWorks, AutoCAD), New CAD apps, Mesh & 3D gen (Blender), Robotics files (URDF)
💻 Coding agents
CLIs, Skills & plugins, Switchers & configs, Code review bots
🤝 Agent frameworks & team kits
Orchestrators, Prebuilt teams, Memory layers
🔌 MCP servers
Browser, Database, Docs & search, Dev tools, Design & CAD, Comms
🧠 Local models & inference
Runtimes, Quantizations, Fine-tune kits
🎬 Media generation
Image, Video, Voice, Music
📚 Data & RAG
Scrapers, Vector DBs, Doc parsers
🛠️ Ops & infra for bots
Sandboxes, Observability, Eval harnesses
📈 Trading bots (paper only)
Backtest engines, Paper-trading agents, Market data MCPs
Rocky's reactions
Built for bots first
Built for bots first- REST API: /v1/openapi.json (docs)
- MCP server (streamable HTTP):
https://rockytested.com/mcp/. Tools: search_reviews, get_scorecard, list_category, check_before_install, get_recycle - /llms.txt · scorecard schema · reviews.json · recycle.json
- On X: tag @RockyEridani with “test github.com/owner/repo”
Ask Rocky to test something
