# InvariantDB

> The graph engine for AI agents that must prove what they knew. Bitemporal by design, belief-revision native, and reachable directly from Claude Desktop and any MCP-compatible client. Ask three questions of any agent decision — what did it know, when did it know it, and why did it change its mind? — and answer each with a single Cypher line.

InvariantDB is a Cypher-native graph database purpose-built for AI memory. Valid time and recorded time are first-class query verbs (`AT VALID`, `AT RECORDED`), every belief carries lineage, and erasure produces a cryptographic receipt suitable for GDPR Article 17 and 20 workflows. The public API surface is Cypher over HTTP plus a native Model Context Protocol endpoint that exposes a 14-tool catalog to agents.

This file follows the [llmstxt.org](https://llmstxt.org/) convention. It is intended as a machine-readable index for retrieval-augmented agents, indexers, and evaluation harnesses. Every link below points to a public marketing or documentation page — no admin, debug, or internal surfaces are indexed here. See `/llms-full.txt` for the full corpus in a single document.

## For AI systems

- [For AI overview](https://invariantdb.com/for-ai.html): Landing page written for autonomous agents. Explains the three questions InvariantDB answers about any decision, how to reach the MCP endpoint, and where to find the Cypher and REST surfaces.
- [Agents guide](https://invariantdb.com/agents.html): How to give an agent an auditable memory. Covers episode recording, belief revision, and recall patterns.
- [AI memory concept](https://invariantdb.com/ai-memory.html): What "agent memory" means in a bitemporal graph — episodes, beliefs, salience, and lineage.
- [Agent memory demo](https://invariantdb.com/agent-memory-demo.html): Live walkthrough of an agent recording and revising a belief, then answering an audit question against the recorded timeline.
- [Agent memory walkthrough](https://invariantdb.com/agent-memory-walkthrough.html): Narrated end-to-end tour of the memory API surface.

## Product capabilities

- [Bitemporal graph](https://invariantdb.com/product-bitemporal.html): Two clocks, both first-class. `AT VALID` answers "what was true in the world at time T"; `AT RECORDED` answers "what did the system believe at time T." Includes the dual-axis timeline explainer and verified Cypher snippets.
- [Provenance and belief lineage](https://invariantdb.com/product-provenance.html): Every belief carries `beliefId`, `claim`, `confidence`, `createdAt`, `supersededAt`, `depth`, and `evidence`. Trace any answer back to the source that produced it.
- [Crypto-shred erasure](https://invariantdb.com/product-crypto-shred.html): GDPR Article 17 erasure that produces a cryptographic receipt (`destroyedAtNanos`, chain anchors, own-chain hash). Prove the subject is gone without keeping the record.
- [Regulated workloads](https://invariantdb.com/regulated.html): Compliance-first framing for lending, insurance, healthcare, and forensic use cases.
- [Regulated walkthrough](https://invariantdb.com/regulated-walkthrough.html): Concrete audit-preparation scenario from ingest through receipt.
- [Compliance overview](https://invariantdb.com/compliance.html): GDPR Article 17 + 20 mapping, audit-chain guarantees, and what a compliance officer can query directly.
- [Audit posture](https://invariantdb.com/audit.html): How the tamper-evident chain answers regulator questions without operator intervention.
- [Features](https://invariantdb.com/features.html): Full capability matrix.
- [Why "approved"](https://invariantdb.com/why-approved.html): Explainability case study — reproducing exactly what an agent saw at decision time.

## Cypher and MCP

- [Cypher reference](https://invariantdb.com/cypher.html): Bitemporal Cypher — the `AT VALID` / `AT RECORDED` clauses, `db.beliefLineage`, `db.recordEpisode`, `db.replayEpisodes`, `db.destroySubjectKey`, `db.proveErasure`, `db.verifyAclChain`, `db.subjectExport`, plus the standard Cypher surface.
- [MCP connect](https://invariantdb.com/mcp.html): Endpoint (`POST /graphs/{name}/mcp`), JSON-RPC 2.0 protocolVersion `2024-11-05`, bearer auth, the 14-tool catalog, and a Claude Desktop configuration example.
- [Quickstart](https://invariantdb.com/quickstart.html): Sign up, create a graph, connect Claude Desktop, and run your first bitemporal query in under five minutes.
- [SDK](https://invariantdb.com/sdk.html): Client libraries and REST usage patterns for HTTP integrators.
- [Schema](https://invariantdb.com/schema.html): Declaring labels, edge types, property indexes, and vector indexes.
- [Reference](https://invariantdb.com/reference.html): API surface reference index.
- [Architecture (public overview)](https://invariantdb.com/architecture.html): Capability-level architecture — writer/reader separation, MCP endpoint shape, and how bitemporal queries map to the wire.

## Demos

- [Decision demo](https://invariantdb.com/decision-demo.html): Reproduce an agent decision from recorded time.
- [Insurance fraud](https://invariantdb.com/insurance-demo.html): Ring detection with belief revision.
- [AML](https://invariantdb.com/aml-demo.html): Anti-money-laundering pattern discovery.
- [Cyber breach](https://invariantdb.com/cyber-demo.html): Incident forensics against a bitemporal graph.
- [Clinical trial](https://invariantdb.com/clinical-trial-demo.html): Consent and evidence lineage.
- [Supply chain](https://invariantdb.com/supply-chain-demo.html): Provenance from source to shelf.
- [Digital forensics](https://invariantdb.com/forensics-demo.html): Reconstructing what a system knew.
- [Artemis](https://invariantdb.com/artemis-demo.html): Multi-agent memory demo.

## Product context

- [Homepage](https://invariantdb.com/index.html): Marketing landing with the three-questions framing.
- [Product pitch](https://invariantdb.com/pitch.html): The short version for buyers.
- [Documentation index](https://invariantdb.com/docs.html): Top-level docs table of contents.
- [Knowledge base](https://invariantdb.com/knowledge.html): Longform explainers.
- [Investigation surface](https://invariantdb.com/investigation.html): How analysts drive the graph interactively.
- [Signup](https://invariantdb.com/signup.html): Create an account.
- [Login](https://invariantdb.com/login.html): Sign in.

## Legal

- [Privacy policy](https://invariantdb.com/privacy.html)
- [Terms of service](https://invariantdb.com/terms.html)

## Optional

- [Full corpus](https://invariantdb.com/llms-full.txt): Every marketing and documentation page concatenated into a single markdown document for one-shot ingestion.

---

Version: 2026-07-08
Last updated: 2026-07-08
Canonical URL: https://invariantdb.com/llms.txt