Enterprise API & AI-Native Integration

Obsidian API & MCP.

Programmatic access to a structured, enriched regulatory data flow. One canonical schema, distributed in every format your stack already speaks: JSON, JSONL, CSV, Parquet, Atom, MCP. Plug it into your data warehouse, your SIEM, or your AI agent.

Talk to sales See it in action

Structured & enriched

Two protocols, one feed.

# 1, subscribe
subscribe feed(framework="CSRD")

# 2, you receive
entry {
   title         "EFRAG publishes ESRS XBRL taxonomy v1.2"
   source        "EFRAG"          official
   framework     "CSRD"
   jurisdiction  "EU"
   industry      "esg"
   relevance     0.94            high
}

# Standard Atom under the hood, plug into any reader, SIEM, or pipeline.

# 1, ask, in plain English
ask "What changed on PFAS this week, EU only, top 5"

# 2, you receive
article {
   title         "ECHA opens consultation on PFHxA derogations"
   source        "ECHA"           official
   published     "2026-05-04"
   jurisdiction  "EU"
   industry      "chemicals"
   relevance     0.91            high
}
# + 4 more, same shape, all enriched

# Native MCP. Works with Claude, Cursor, and any compliant client.

feed healthy API p50 312ms | uptime 99.95% (90d) | # Publication-to-feed latency: 30 min to 24 h^*, depending on source.

Data flow

A clean stream of structured regulatory data.

One enriched canonical record per regulatory event, distributed in every format your stack already speaks. From parser to your data warehouse, your SIEM, your AI agent.

Regulators EU, US, CH...

Official journals EUR-Lex

Standard-setters ISO, EFRAG

Agencies ECHA, EBA

Native feeds RSS, sitemap

obsidian.pipeline v3.2

Parse & normalize HTML, XML, PDF, OCR. One canonical schema. schema
Deduplicate & version Stable IDs, hashed payloads, full diff history. idempotent
Classify Industry, framework, jurisdiction tagging. tagged
Score & enrich Relevance, severity, references, citations. scored
Validate Schema checks, integrity hash, provenance kept. audited

REST / JSON api

JSONL stream ndjson

CSV / Parquet batch

Atom / RSS feed

Webhooks push

MCP server ai

1, canonical schema

One record, one source of truth

6+ formats

JSON, JSONL, CSV, Parquet, Atom, MCP

99.95%

API uptime, trailing 90 days

SOC 2 aligned

Audit logs, signed payloads, retention

^* Publication-to-feed latency typically ranges from 30 minutes to 24 hours, depending on the source and its publication mode (push, pull, or scheduled crawl).

Path 1, RSS

For your infrastructure.

Standard Atom feeds with Obsidian's enrichment fields. Plug them into your SIEM, GRC platform, internal data lake, or any feed reader. Zero integration cost.

Bearer auth Per-framework feeds Auto-refresh

Path 2, MCP

For your AI assistants.

Native MCP server exposing structured tool calls. Connect Claude, Cursor, or any compliant client and query the feed in natural language. No glue code.

Tool calls Natural language Any MCP client

Both protocols, same canonical schema. Pick the path that fits today, switch on the other when your stack evolves.

Request access

Ready to plug regulatory intelligence into your stack?

Tell us about your industry, your jurisdictions, and your stack. We'll set you up with the right path, the right scope, and the right SLA.

Contact sales See all solutions

Request a Demo

Obsidian API & MCP.

For your infrastructure.

For your AI assistants.

Ready to plug regulatory intelligence into your stack?