RavenGet a demo

One platform.
Three products.

Raven is a single, composable platform for institutions that need to reason and act on complex environments. Adopt one product or run the whole stack — sense, decide, act.

Get a demoRead the docs
Products
Muninn — the memory layer
// 01Live
Muninn

The memory layer.

Memory that remembers what others miss. Muninn ingests every signal — telemetry, documents, the open web — and keeps it queryable forever, with provenance attached.

Read more
Huginn — the thought layer
// 02Live
Huginn

The thought layer.

Thought that pressure-tests the answer. Huginn turns Muninn's evidence into decisions — with adversarial review, scenario simulation, and audit on every step.

Read more
Sleipnir — the eight-legged horse
// 03Live
Sleipnir

The eight-legged horse.

Odin's mount. The fastest path from thought to outcome. Sleipnir carries the decision Huginn signed to wherever it needs to fire — under your policy, under your authority, with every step replayable on demand.

Read more
How the platform works

Sense, decide, act.
Bound by one substrate.

The same data, identity, audit, and policy layer underpins all three products — so what enters Muninn can be reasoned about in Huginn and acted on by Sleipnir without losing context.

  1. Step 01Sense

    Muninn turns signal into structure.

    Streaming and batch ingestion across warehouses, message buses, telemetry, documents, and the open web. Vector, graph, and structured indexes share one query plane. Provenance is enforced at write time.

    • 120+ connectors
    • Hybrid retrieval
    • Anomaly detection
  2. Step 02Decide

    Huginn turns structure into decisions.

    Analysts assemble cases on a structured canvas. Reasoning models pressure-test conclusions, simulate alternatives, and surface counter-evidence. Every step is auditable, every conclusion ships with the chain that produced it.

    • Evidence graphs
    • Adversarial review
    • Signed exports
  3. Step 03Act

    Sleipnir turns decisions into outcomes.

    Policy-enforced workflows fire on events. Agents act, humans approve where it matters, and every step is replayable. Native integrations land output in Slack, PagerDuty, ServiceNow, or any HTTP endpoint.

    • Policy as code
    • Human gates
    • Replayable audit
Stack at a glance

A coherent stack from raw signal to operational outcome.

01SenseMuninnA live, queryable picture of your environment built from every source you care about.
02DecideHuginnA workspace where evidence becomes decision — with reasoning models, simulation, and audit.
03ActSleipnirAutonomous workflows that act on policy with full human oversight.
Integrations · 120+

Connect to everything you already run.

Raven plugs into the systems your business runs on — and the ones your regulators ask about. Native connectors for the most common stacks, plus a typed HTTP surface for everything else.

Data warehouses
  • Snowflake
  • Databricks
  • BigQuery
  • Postgres
  • Redshift
Cloud & infra
  • AWS
  • Azure
  • GCP
  • OVHcloud
  • Scaleway
Messaging & ITSM
  • Slack
  • Microsoft Teams
  • PagerDuty
  • Opsgenie
  • ServiceNow
Identity
  • Okta
  • Azure AD
  • Google Workspace
  • JumpCloud
  • Auth0
Source systems
  • SAP
  • Oracle
  • Salesforce
  • Workday
  • Custom HTTP
Document repos
  • SharePoint
  • Confluence
  • Drive
  • Notion
  • S3
In production

The platform is the first thing in years that respects how analysts actually work. Provenance everywhere. No more hunting for the source.

Head of IntelligenceIntelligence · EU energy operator
Frequently asked

Questions we get from institutions.

How is Raven different from a data warehouse, BI tool, or workflow engine?

+
Each of those does a slice. A warehouse stores. BI visualizes. A workflow engine orchestrates. Raven binds them: Muninn ingests and indexes, Huginn reasons and decides, Sleipnir executes — under one identity, audit, and policy substrate. The output is an accountable decision, not a query result.

Can we adopt one product without the others?

+
Yes. The most common entry point is Muninn, deployed alongside an existing warehouse and analytics stack. Customers add Huginn and Sleipnir as the use cases mature. The pieces are designed to compose, not to require all-or-nothing.

Where is data stored and processed?

+
EU/EEA by default across all deployment modes. Choose a specific region (Ireland, Frankfurt, Madrid, Stockholm), run inside your own EU cloud account, or deploy on-premises. Transatlantic data transfer paths are off by default.

How does Raven handle the EU AI Act?

+
Huginn and Sleipnir are designed against the AI Act's high-risk obligations: human oversight (Article 14), transparency (Article 13), and documentation. Audit, signed reasoning chains, and explicit human approval gates are first-class features, not configurations.

How long does deployment take?

+
First production use case typically lands in 6–10 weeks. The first connector and first decision-grade workflow ship in the same engagement. Forward-deployed engineers work alongside your team for the first quarter.

What happens to our data if we leave Raven?

+
Full export of every observation, case, and audit record in open formats. Customer data is purged from Raven systems within 30 days of termination. No vendor lock — by contract.

Build with the full stack.

Get a demo Documentation