pipr
Project Intelligence & Planning Runtime
pipr is an experimental, open-source planning companion for solo founders and small development teams.
It helps turn current intent into concrete, explainable execution plans by grounding LLM-based planning in explicit project context and accumulated decisions.
pipr is not a task tracker or project management system.
It sits before execution tools like GitHub and helps you decide what to do next — and why.
For a detailed explanation of pipr’s planning model, terminology, and design rationale, see
docs/wiki.md
What problem pipr solves
Early-stage projects move quickly, assumptions change often, and most planning context lives only in a developer’s head.
This leads to:
- Repeated analysis
- Vague or redundant task lists
- Loss of reasoning behind decisions
- Cognitive overload when planning the “next step”
pipr helps by:
- Reducing planning and triage cognitive load
- Turning fuzzy goals into small, actionable task proposals
- Preserving why decisions were made, not just what was done
- Accumulating planning context over time so you don’t have to repeat yourself
Core idea
Planning quality depends more on structured context than on clever prompting.
pipr makes planning context explicit, persistent, and inspectable — then uses an LLM to reason within those constraints.
What pipr deliberately does not do (yet)
- Act as a full project management tool
- Track execution state or task progress
- Automatically optimize or rank proposals
- Learn preferences or adapt behavior autonomously
- Read GitHub issue state during planning
These may come later — but only if they prove necessary.
Intended users
- Solo founders iterating toward product–market fit
- Small dev teams without a dedicated PM
- Developers who want better planning without heavy process
pipr is especially useful when:
- You’re context-switching frequently
- You want to avoid re-thinking the same decisions
- You care about why something was planned, not just what
Tech
- Backend: Fastify + tRPC
- Frontend: React + Vite
- Database: Postgres + Prisma
- LLMs: Local (Ollama) or hosted providers
- Context storage: Plain-text authoritative project context
Status
pipr is under active development and is being used to plan its own evolution.
The planning model is intentionally minimal and extensible.
Expect breaking changes.
Planned Features
-
Context retrieval at scale (RAG)
- Semantic retrieval over accumulated planning signals
- Used only when context size exceeds what fits in a single planning frame
-
Additional planning signals
- Strategic intent (longer-term direction that outlives individual goals)
- Explicit provenance signals (why a signal exists, not just what it says)
-
Completed work inference
- Populate
completed_worksignals from external execution systems (e.g. GitHub issue closure)
- Populate
License
Apache License 2.0
See LICENSE.