Dome Systems

Integrations

Everything under the Dome

The models, tools, identity providers and agent runtimes Dome governs. Pick a kind, or search.

If it speaks the standard, it works

MCP for tools, OpenAI- and Anthropic-compatible APIs for models, OIDC for identity.

AI21Providers
Amazon BedrockProvidersAmazon Bedrock AgentCoreRuntimesAnthropicProvidersApifyAgent servicesApolloAgent servicesAtlassianMCP serversAuth0IdentityAzure OpenAIProvidersBrowserbaseAgent servicesCerebrasProvidersClaude Agent SDKRuntimesClaude CodeClientsClaude FableModelsClaude HaikuModels

Operations integrations

Where audit and events go

Agents call what sits under the Dome. Your security and operations tools hear about every call. Stream the audit trail to them, or send events to any of these presets or your own webhook.

Audit export

Stream the full audit trail, in OCSF or CEF, to the store your security team searches.

  • Amazon S3
  • Google Cloud Storage
  • Datadog
  • Grafana Loki

Observability and SIEM

Send events to the dashboards and detection rules your teams already run.

  • Datadog
  • Splunk
  • Elasticsearch
  • Sumo Logic
  • New Relic
  • Grafana Loki
  • Dynatrace
  • Honeycomb
  • Axiom
  • Better Stack
  • CrowdStrike Falcon LogScale
  • Google SecOps
  • Panther

Incident and chat

Page the on-call engineer on a denial spike, or post an agent's suspension to the team channel.

  • PagerDuty
  • incident.io
  • Jira Service Management
  • ServiceNow
  • Slack
  • Microsoft Teams
  • Discord

Automation and delivery

Trigger playbooks and workflows from Dome events, and run the dome CLI in your pipelines.

  • Tines
  • Torq
  • Zapier
  • Make
  • n8n
  • Jira Automation
  • Airtable
  • Azure Event Grid
  • GitHub
  • GitLab CI

Events

Audit is the record. Events are the reflex.

Every governed action lands in the audit trail — complete, immutable, retained, and carrying the full evidence. Events are the other half of that: selected facts, projected to a public contract, and delivered to the systems you already run. The split is deliberate. It keeps the audit schema free to evolve without freezing it into a public API, and it means the payload leaving Dome is public-safe by construction.

Signed, at-least-once delivery

Every event is a signed POST against a delivery ledger, with exponential backoff on failure. Verify the signature at your endpoint and dedupe on the event id.

A public, versioned contract

Events are published against a generated catalog — 55 types, each with a version, a maturity stage, and a JSON Schema. Only registry-approved fields are projected, so a webhook endpoint cannot become a PII sink.

Curated groups, not a type list

The create flow offers bundles — agent lifecycle, governance changes, security signals, incidents — so a destination can be subscribed to what it cares about without assembling event types by hand.

Destinations by preset

Point at an off-the-shelf destination by picking it and pasting a credential, rather than hand-configuring a raw webhook. Each preset carries its own form, endpoint shape, and recommended bundle.

Delivery history and replay

See exactly what was sent to each destination and what came back. Replay a failed delivery, or replay everything that failed since a point in time.

Per-destination circuit breaking

A failing destination is isolated and retried on its own schedule. One broken endpoint cannot slow or stall delivery to the healthy ones.

In practice

Subscribe a destination

Pick a preset and paste its credential, or point at your own endpoint. Every delivery is signed, and failed ones replay.

Off-the-shelf destination

$ dome webhooks providers list
 
$ dome webhooks subscriptions create sec-ops \
--provider pagerduty \
--credential routing_key=$PD_ROUTING_KEY \
--event-type access.denied \
--event-type agent.suspended

Custom endpoint, then replay

$ dome webhooks subscriptions create siem \
--url https://siem.example.com/dome \
--event-type access.denied
 
$ dome webhooks deliveries list
$ dome webhooks deliveries replay-failed --since 24h

Next steps

Talk with our FDE team

Our forward deployed engineers work with your platform team to get your agents into production and under control: the first one governed on your own systems, and a pattern your teams can repeat for every agent after it.