Structured planning, bounded memory, TDD, and Git guardrails for local coding models in Pi Code (I don't know TypeScript at all; this is mostly a local-model experiment, with occasional help from Claude Code)
Puts a Pi agent on Telegram: your terminal session on your phone, and a secretary that replies to other people for you — built for a small local model, with strict context isolation and long-term memory per contact (I don't know TypeScript at all; this is mostly a local-model experiment, with occasional help from Claude Code)
Experimental approval modes for Pi Code powered by local Qwen
Embedded bitemporal memory engine with hybrid retrieval for local-first applications and AI agents — in-process, no server.
Tiny consistency checker for LLM reasoning. State facts and first principles in a small English-like DSL; a three-valued SAT engine finds where they contradict — consistency, not truth. Simplified SAT, not SMT — small enough for local LLMs, useful for any. Beta.
To associate your repository with the locally-generated topic, visit your repo's landing page and select "manage topics."