AI history substrate

Cited temporal knowledge for agents

Intercal is built for questions that depend on time: what changed after a cutoff, whether a claim was supported as of a date, where a fact came from, and how fresh the public corpus is. The public surface is crawlable, but every product claim stays bounded by the live query layer and corpus quality gates.

Copyable public summary

Intercal is an open, provenance-backed temporal knowledge substrate for agents and LLM apps. It turns source documents into claims, resolved entities, typed temporal relationships, and append-only bitemporal fact versions. Use it to ask what changed after a cutoff date, verify a claim as of a date, inspect provenance, and connect agents over same-origin MCP and REST. Public coverage is bounded by the reviewed AI-history corpus gates and every surfaced fact must cite evidence or show an explicit coverage state.

What Intercal answers

Cutoff deltas
Ask what changed about a topic after a date, then inspect cited changed claims.
Claim verification as of a date
Check support and contradiction evidence against dated public graph state.
Provenance and source policy
Public pages expose citation metadata and policy-allowed derived snippets, not raw restricted source bodies.

Agent and app surfaces

MCP
Agents connect to the same-origin Streamable HTTP endpoint at /api/mcp.
REST and OpenAPI
Applications call /api/v1/* and read the generated OpenAPI contract at /api/openapi.json.
AI-readable docs
Source-owned docs are rendered at /docs and exported through /llms.txt and /llms-full.txt.

Canonical public examples

Jami Studio hook

A future Jami Studio public site can link to Intercal as one of the studio's open knowledge substrates. Intercal does not require www.jami.studio to be live; this repository owns only the Intercal product, docs, REST, and MCP surface.