inform
Background Corpus Management
Automate external knowledge base maintenance
Automated Orchestration
Coordinates fetch/extract (inform) and package/index (catalog) workflows for seamless corpus updates. Perfect for hooks and background automation.
Git-Friendly Configuration
Keep corpus configuration and manifests in git. Track your sources and metadata while actual documents live on shared mounts for team collaboration.
Configurable Timeouts
Prevent hanging on problematic sources with per-source timeout configuration, retry limits, and continue-on-error behavior.
Error Logging
Comprehensive error tracking with detailed logs written to ~/.hyphn/logs/inform-sync.log
for debugging and monitoring.
Multi-Agent Sharing
Share corpora across multiple agents via shared mounts. Configuration in git, content on external storage—best of both worlds.
Flexible Phasing
Run specific phases with --only pull|index|export for granular control over
corpus updates.
Orchestration Workflow
Inform coordinates corpus maintenance across multiple sources
Fetch/Extract
Inform pulls from sources
Writes to corpus content
Package/Index
Catalog processes content
Writes indexed artifacts
Manifests
Update configuration
Commit to git
# Complete corpus sync inform sync --all --mirror --print-cmds # Status check inform status
Installation
Set up corpus management
Bun Package
Run from monorepo for full ecosystem access
Quick Start
Get your corpus configured and syncing
Initialize Configuration
Sync Corpus
External vs Internal Knowledge
Inform manages external docs, Reflect manages internal history
inform (External)
Manages system's external knowledge base: documentation, APIs, tutorials fetched from external sources. Updates source info, catalogs content, handles cleaning.
reflect (Internal)
Focuses on system history to create and curate internal knowledge base from previous experiences, learnings, and session outcomes.
Advanced Configuration
Fine-tune timeouts and error handling
Timeout Configuration
Per-source timeout in milliseconds (default: 5 minutes)
Retry Strategy
Maximum retry attempts per source (default: 2)
Error Handling
Continue syncing other sources if one fails (default: true)
# Strict sync with short timeout FWD_CORPUS_TIMEOUT=60000 FWD_CORPUS_CONTINUE_ON_ERROR=false inform sync # Lenient sync with extended timeout FWD_CORPUS_TIMEOUT=600000 FWD_CORPUS_RETRIES=3 inform sync
Environment Variables
Configure behavior via environment
Ready to Automate Corpus Management?
Keep your external knowledge base fresh with automated workflows