This directory is the canonical home for Chronicle's technical documentation. Start with the repository Quick Start Guide for installation and day-to-day operation, and use AGENTS.md for development conventions.
- Project overview: components, deployment topology, and repository layout
- Testing and coverage: fast Python lanes, coverage reports, and integration tests
- Release process: candidate validation, TestFlight, publication, and verification
- Audio pipeline: session, transcription, and memory flow
- Multimodal memory: observations, event discovery, evidence retrieval, and durable memory
- Initialization system: setup wizard and service orchestration
- Fleet updates: how nodes learn about and apply code updates
- ScreenPipe capture nodes: local capture, Chronicle ingestion, desktop controls, and logs
- Podman: rootless containers, GPU access, and engine migration
- SSL certificates: HTTPS setup and certificate trust
- Authentication: user identity, JWTs, and protected endpoints
- Memory system: agentic Markdown vault and retrieval
- Audio durability: raw-audio write path and its state machine
- Data archive and memory rebuild: full export/import and clean vault reconstruction
- Plugin configuration: configuration and secret boundaries
- Plugin development: plugin lifecycle and APIs
- Frontend UX review notes: Chronicle-specific UI conventions
Point-in-time investigations. Each states its own date and scope; they are kept for their evidence, not as descriptions of the current system.
- Codex vault evaluation: Codex CLI memory executor compared against the direct tool-calling vault agent
- Memory rebuild audit: read-only audit of a live database and its generated vault
- Vault iteration plan: follow-up to the rebuild audit
Design plans (docs/plans/) and the screen-memory research reports
(docs/research/) are kept locally and are not part of the repository.
Service-specific instructions remain beside their implementations, such as
extras/speaker-recognition/ and
extras/tts/.