Your workflows have
no memory.We fix that.
Smriti gives ops teams, AI agents, and SaaS platforms a bi-temporal event ledger — auditable, queryable, and live in 3 API calls.
Vector sludge
Traditional RAG dumps unstructured text into a vector DB with no time order. Old facts and current state become indistinguishable.
No temporal tracking
Workflows that track deals, deployments, or contracts need to know when something changed — not just that it did.
Silent overwrites
Without a supersession engine, correcting a fact silently destroys history. Audits become impossible; conflicts go undetected.
Three calls. Permanent memory.
Feed any text
Send raw data from CRMs, chats, emails, or git commits. Smriti accepts completely unstructured text.
AI extracts structure
Our pipeline automatically extracts Subject-Verb-Object (SVO) events and stores them in PostgreSQL with dual pgvector indexing.
Query anything
Search naturally. Smriti returns exact, structured JSON events in ~80ms, enabling your agents to reason temporally.
Connect
Three ways to integrate.
Start with a terminal. Route to your own database. Give your AI tools direct access. Each method is independent — use one or all three.
Any language, any tool
Direct API
The fastest way to start. One curl, one result.
Send raw text to /ingest. Get structured, time-stamped SVO events back from /query. No SDK, no setup — works from any terminal, language, or HTTP client.
# Store an event
curl -X POST https://spy9191-chronos-api-backend.hf.space/ingest \
-H "X-API-Key: chrn_your_key" \
-d '{"source_id": "crm", "events": [{"text": "Acme signed $50k contract"}]}'
# Recall it
curl -X POST https://spy9191-chronos-api-backend.hf.space/query \
-H "X-API-Key: chrn_your_key" \
-d '{"query": "What happened with Acme?"}'Self-hosted · Supabase
Bring Your Own Database
Your data never touches our storage.
Add one header. All extracted events, embeddings, and SVO tuples route directly to your own Supabase (PostgreSQL + pgvector) project. The pipeline runs identically — your infra, your control.
curl -X POST https://spy9191-chronos-api-backend.hf.space/ingest \
-H "X-API-Key: chrn_your_key" \
-H "X-Supabase-Url: postgresql://postgres.xxxx:pw@\
aws-0-us-east-1.pooler.supabase.com:5432/postgres" \
-d '{"source_id": "crm", "events": [{"text": "Acme signed $50k"}]}'Claude · Cursor · VS Code
Model Context Protocol
Your AI tools gain memory. No code to write.
Smriti ships a built-in MCP server. Configure it once and Claude Desktop or Cursor get full read/write access to your event ledger mid-conversation — remember, recall, forget, timeline.
# Install pip install -r mcp/requirements.txt # Set your key export SMRITI_API_KEY="chrn_your_key" # Run the server python -m smriti.mcp
How it looks in code
# Feed an event
httpx.post("/ingest", json={"text": "Acme signed a contract"})
# Query memory
httpx.post("/query", json={
"query": "What happened with Acme?"
})
# → { subject: "Acme", verb: "signed"... }
Explorer
- → 10,000 events/month
- → Standard query speed
- → Community support
- → 3 connected tools
- → 1,000 orchestrations/month
No credit card required
Give your workflows memory.
An auditable event ledger for every team. Free to start, live in 5 minutes.
Get Free API Key →Free tier · 10,000 events/month · No card required