MCP Server | Revelara

Revelara exposes a research server over the Model Context Protocol (MCP). Connect it to your coding agent and the agent can ground reliability decisions in evidence: the public incident corpus, your organization’s risk register, the reliability controls catalog, distilled SRE knowledge, your service catalog, and the knowledge graph that links them.

All tools are read-only, and every result is scoped to your organization.

Endpoint


URL	`https://api.revelara.ai/mcp`
Transport	MCP streamable HTTP
Auth	Revelara API key (Bearer) or OAuth

Authentication

Two options, depending on the client:

API key. Create a key at app.revelara.ai under Settings > API Keys and send it as a Bearer token. This works with Claude Code and any MCP client that supports custom headers.
OAuth. When you add Revelara as a custom connector on claude.ai, authentication happens through an OAuth login. No key handling required.

Connect from Claude Code

claude mcp add --transport http revelara https://api.revelara.ai/mcp --header "Authorization: Bearer pk_live_YOUR_KEY"

Replace pk_live_YOUR_KEY with an API key from Settings > API Keys.

Tools

Group	Tools	Use for
Incidents	`search_incidents`, `get_incident`, `get_incidents_batch`, `find_similar_incidents`, `get_incident_stats`	What has failed before, and how. Search the corpus, pull full incident bodies, find incidents similar to a scenario, get counts and distributions.
Knowledge	`search_knowledge`, `retrieve_memories`	Distilled facts and procedures, higher signal than raw incident text. `retrieve_memories` recalls prior conversation context.
Risks	`search_risks`, `get_risk_context`	Your organization’s open exposure. `get_risk_context` returns a risk with its mapped controls, evidence, and service context.
Controls	`search_controls`, `get_control`, `analyze_control_safety`	What to implement. `analyze_control_safety` gives an STPA-inspired view of a control’s failure modes.
Graph	`list_entities`, `explore_graph`, `get_entity_relationships`, `compare_incident_groups`	Trace causal chains across the corpus, discover filter axes, contrast two incident populations.
Catalog	`search_service_catalog`, `get_service_catalog`	Discover services and look up a service’s tier, owners, stack, and dependencies.

Conventions

Wrap a phrase in double quotes for exact-phrase matching, e.g. "cache stampede".
Cite results by their stable IDs: incidents by short_name, risks by R-XXX, controls by RC-XXX.

Other ways to connect

CLI. brew tap revelara-ai/tap && brew trust revelara-ai/tap && brew install --cask rvl installs the rvl CLI for risk scans, evidence submission, and shell-composable queries against the same data.
Claude Code plugin. rvl plugin install claude installs Revelara’s slash commands and skills into Claude Code.

Access

The MCP server requires a Revelara account. Sign up at app.revelara.ai to get started.