Tunes a local vLLM or llama.cpp server by measuring one flag change at a time
- Open source
- 4 Hacker News points
- Developer tools
Tunes a local vLLM or llama.cpp server by measuring one flag change at a time
Blind-graded LLM evals to find which model and prompt score best on your own tasks
Open source, local AI workspace with custom interactive UIs for your own agents
AI chat over your team's code snippets library, with codebase indexing and multiple LLMs
Claude Code add-on that estimates task time in measured AI working minutes
Self-hosted proxy that logs token use, cost and latency for LLM API calls and enforces key
Self-hosted OpenAI-compatible gateway that load-balances multiple model upstreams
AI that explores your website, checks it for broken pages and accessibility, and drafts
Experimental browser IDE for steering LLM text generation in realtime

Learns from past agent sessions to cut coding-agent token costs
Rust/GPU tool that rotates an LLM's output matrix to keep correct answers and repair wrong

Ruby library with one API for 23 AI providers, tools and Rails conventions
The closest alternatives to Werkbank are Tensward, PromptJury, Lynx, codesnippets and AI Time Estimator.
Yes. Tensward, PromptJury, Lynx have a free or open-source option.