Show HN: Project Journal – Give AI coding assistants persistent memory (github.com)

Project Journal is a lightweight, CC0 markdown template that gives AI coding assistants persistent, human-curated memory across sessions. Drop PROJECT_JOURNAL.md into your repo root, fill sections like Team & roles, Project overview, Tech stack decisions, Completed features, Design rationale, Session logs, Current status and Next steps, then prompt your assistant to "Read PROJECT_JOURNAL.md" at the start of a session and to "Update PROJECT_JOURNAL.md" at the end. The repo includes a downloadable PROJECT_JOURNAL_TEMPLATE.md and a real-world example from Vibe CMS to show how to structure entries and session logs. For the AI/ML community this is a pragmatic workaround for stateless LLM sessions: it preserves context, rationale, and progress without building specialized memory infrastructure. Technical implications include faster iteration (no repeated onboarding), better traceability of design choices, easier collaboration and automated prompts (e.g., “Summarize this session for the journal”), and version-controlled memory when you commit journal updates alongside code. It’s especially useful for solo devs, teams, and students working with assistants like ChatGPT/Claude/Copilot — a simple, prompt-driven pattern that complements model capabilities by making project history explicit and machine-readable.