Self-hosted: Free, MIT licensed. Hindsight Cloud: Pay-as-you-go with free credits to start. Retain (store memories) $10.00 per million tokens, Recall (retrieve memories) $0.75 per million tokens, Reflect (synthesize patterns) $0.05 per call, Iris Extract (structured extraction) $7.50 per million tok
Free: 1,000 credits/month, no card; Builder: $49/month for 10,000 credits; Pro: $199/month for 50,000 credits; Scale: $799/month for 250,000 credits. Enterprise with custom terms available. Credits cost approximately $0.0032-$0.0049 per page, $3.20-$4.90 per 1,000 pages, or about $0.19-$0.29/hour of
Key features
Per-user persistent memory with preferences, history and decisions kept separate
Cross-session persistence allows context to survive weeks between interactions
Fast parallel retrieval in under 100ms through four strategies running simultaneously
Works with any LLM without losing learned context when swapping models
Learns from mistakes and corrects behavior on repeated problems
Automatic pattern detection synthesizes individual facts into consolidated knowledge
Conflict detection preserves both old and new states with timestamps when facts change
Integrates with 20+ coding agents including Claude Code, Cursor, and Copilot
Scrape: Convert any web page into clean markdown or structured JSON
Crawl: Fetch every page on a site from a single link and receive results as they are ready
Search: Query the web with top results already read and parsed
Browsers: Real browser instances in the cloud controlled via Playwright, Puppeteer or SDK
Agents: Autonomous task execution in a real browser with plain English descriptions
Session recordings: Every browser session is recorded for debugging and verification
Monitor: Watch pages on a schedule and receive notifications when content changes
Saved logins: Store credentials securely for agent runs requiring authentication
MCP integration: Connect to Claude Code, Codex or Cursor with one command
Real browser handling: Sites treat agents as humans so pages load and sessions persist
What makes it different
94.6% retrieval accuracy on LongMemEval-S, beating next-best system at 74.0%
MIT licensed open source with no usage limits, telemetry or restrictions
Reflection layer automatically builds judgment and mental models over time without manual tagging
Entity resolution handles hundreds of entities with 99.5% query reduction
Vectorize
Outcrawl