Skip to content
Skip to content
Open District Sign in

The foundation for people and AI to understand, remember, and act.

Open District connects public data, enterprise systems, personal context, memory, reasoning, and governed action in one secure platform.

Build on Open District See how it works
  • Public data
  • Enterprise data
  • Personal context
  • Memory
  • Reasoning
  • Prediction
  • Governed action
How data becomes governed action
  1. Reality Public + enterprise + personal data
  2. Memory Event store — every record retained
  3. Reasoning Query, search, and entity graphs
  4. Prediction Direction — see status section
  5. Action Governed, human-approved tasks
  6. Permissions and audit at every step

Governance (permissions, audit) applies at every stage — see the governance section.

One system for reality, context, and action.

Open District is a platform where people and AI systems can work with trusted context. It connects external data, internal business systems, and personal context into one governed environment.

Public truth

Regulatory notices, legislation, filings, and open datasets — ingested on schedule, queryable with citations back to the original source.

Enterprise truth

Your systems of record — REST APIs, databases, business tools — connected through governed connectors that respect source permissions.

Personal context

Notes, documents, and working context, held privately per user and never mixed into shared knowledge without explicit permission.

Connect the systems where work happens.

Every source is a governed connector: scheduled syncs, schema-on-write to the event store, and citations from every result back to the originating record.

Enterprise systems

  • REST APIs (cursor / offset / API-key / OAuth) Available now
  • Webhooks (inbound) Available now
  • JSON pull (scheduled fetch) Available now
  • GitHub Available now
  • Salesforce (via custom API integration) Available now
  • ServiceNow (via custom API integration) Available now
  • Workday (via custom API integration) Available now
  • File import Available now
  • Jira Platform direction
  • Snowflake Platform direction
  • Postgres (direct) Platform direction

Public data

  • Federal Register (regulatory notices) Available now
  • GovTrack (U.S. legislation) Available now
  • FEC (campaign finance) Available now
  • NYC 311 (service requests) Available now
  • NOAA (space weather + climate datasets) Available now
  • Threat intel (URLhaus, ThreatFox, CISA advisories) Available now
  • Zenodo / OpenLibrary / iNaturalist (research + open data) Available now
  • U.S. Census Platform direction
  • SEC filings (EDGAR) Platform direction
  • Court opinions Platform direction

Full catalog: the in-app Sources page lists all 30+ live connector templates.

What Open District helps you do.

1. Understand

Available now

Ask questions in SQL or natural language over every connected source. Federated queries span Postgres and ClickHouse; semantic search covers the full event history.

2. Remember

Available now

An append-only event store retains every ingested record with its source, timestamp, and payload — the basis for citation-backed answers.

3. Reason

Available now

Entity graphs connect records across sources — find how two entities relate, with each hop cited to the event it came from.

4. Predict

Platform direction

Forecasting and anomaly detection over stored context. Platform direction — not built yet.

5. Act

In development

Turn conclusions into governed tasks with human approval before anything executes. The pipeline engine runs scheduled jobs today; the approval loop is platform direction.

6. Control

Available now

Per-tenant isolation (row-level security), scoped API keys, read-only query gates, and a full audit log of what was asked and answered.

From signal to action.

A regulatory change lands in the Federal Register. Open District catches it, connects it to your records, and carries it through to a governed response.

  1. Regulatory notice published
  2. Open District identifies it
  3. Matches to records
  4. Finds affected accounts
  5. Summarizes impact
  6. Recommends task Target flow — approvals in development
  7. Drafts response Target flow — approvals in development
  8. Human approves Target flow — approvals in development
  9. Logged with sources

Steps 1–5 and 9 run today. Steps 6–8 ship with the approvals loop (in development).

Built for control and accountability.

Private by default. Governed by design.

Permissions

Available now

Every query runs under a tenant context enforced by row-level security. API keys carry explicit scopes; users hold role-mapped permissions.

Privacy

Available now

Tenant data is isolated at the database layer, not just filtered in the app. Personal context stays per-user. Private by default.

Approvals

In development

High-impact actions require a human sign-off before execution. The workflow surface is in development; the gate points are defined.

Audit

Available now

Every query and result is recorded to an append-only audit log with actor, scope, and timestamps — queryable like any other table.

Built for people and machines.

Open District speaks the Model Context Protocol. Point any MCP client at the endpoint below and your agent gets governed access to the same query, search, and audit surface the web app uses — under an API key with explicit scopes.

Primitives

  • Query run_sql — governed read-only SELECT Available now
  • Search unified_search, vector_search Available now
  • Memory the append-only event store Available now
  • Subscriptions standing queries that notify Platform direction
  • Actions governed, approved write-backs In development
  • Audit audit_log table + tool-call records Available now

Full tool reference: the in-app MCP documentation (sign in → MCP).

MCP client configuration

{
  "mcpServers": {
    "opendistrict": {
      "url": "https://app.opendistrict.org/mcp",
      "headers": {
        "Authorization": "Bearer <your-api-key>"
      }
    }
  }
}

Streamable HTTP transport. API keys are issued per tenant with explicit scopes (admin → API Keys). Requests are subject to the same governance and audit as the web app.

What is available now.

This list is verified against the running system. Anything not built is labeled as in development or platform direction — nothing on this page claims a capability the platform does not have.

Available now

  • Event store — append-only, tenant-isolated (Postgres + ClickHouse mirror)
  • 33 source connector templates (REST, OAuth, webhooks, 17 public datasets)
  • Federated SQL queries with a read-only gate and table allowlist
  • Semantic + lexical search (Qdrant vectors, 1024-dim)
  • Entity graph with cited paths between records
  • Pipeline engine — scheduled DAG jobs with sandboxed WASM UDFs
  • MCP server — 10 tools (query, search, schema, catalog, topology, metrics)
  • Audit log — append-only record of queries and results
  • SSO (Zitadel OIDC) + PASETO sessions + scoped API keys
  • In-app notifications (in-app delivery only today)

In development

  • Approvals — human sign-off workflow for high-impact actions
  • Action loop — governed write-backs to connected sources
  • Subscriptions — standing queries that notify on change
  • Prediction — forecasting and anomaly detection
  • Email/SMS notification delivery
  • U.S. Census, SEC EDGAR, court opinions connectors
  • Jira, Snowflake, direct Postgres connectors
  • Billing scaffold (schema landed; not wired)

Start with one source, one team, or one agent.

Build on Open District Talk to us

Clear for people. Useful for business. Ready for AI systems.