HoneyHive MCP server
49 operations0 webhooks
AI agent observability and evaluation platform for tracing, evaluating, monitoring, and improving LLM applications.
Try HoneyHive with your agent
Connect HoneyHive, then paste this into Claude, ChatGPT or Cursor. For ideas from other apps, see the use cases.
Using Corsair, show me what you can do in HoneyHive on my account, then run one read-only example and show me the result.
HoneyHive tools and triggers
Tools are what your agent can do in HoneyHive through Corsair's MCP server. Triggers are the HoneyHive events a routine can start from.
Experiments
- CompareRead-only
Compare two evaluation runs
experiments.compare
- Compare EventsRead-only
Compare events between two evaluation runs
experiments.compareEvents
- Create RunWrite
Create an evaluation run in a project
experiments.createRun
- Create Run LegacyWrite
Create an evaluation run (legacy)
experiments.createRunLegacy
- Delete RunDestructive
Delete an evaluation run
experiments.deleteRun
- Get MetricsRead-only
Get metric events for an evaluation run
experiments.getMetrics
- Get RunRead-only
Get evaluation run details
experiments.getRun
- Get Run LegacyRead-only
Get evaluation run details (legacy)
experiments.getRunLegacy
- Get Run SchemaRead-only
Get event schema for a single run
experiments.getRunSchema
- Get Runs SchemaRead-only
Get event schema across runs in a project
experiments.getRunsSchema
- Get SummaryRead-only
Get summary statistics for an evaluation run
experiments.getSummary
- List RunsRead-only
List evaluation runs in a project
experiments.listRuns
- List Runs LegacyRead-only
List evaluation runs (legacy)
experiments.listRunsLegacy
- Update RunWrite
Update an evaluation run
experiments.updateRun
Datasets
- Add DatapointsWrite
Add datapoints to a dataset
datasets.addDatapoints
- CreateWrite
Create a new dataset in a project
datasets.create
- Create LegacyWrite
Create a new dataset (legacy)
datasets.createLegacy
- DeleteDestructive
Delete a dataset and its datapoints from a project
datasets.delete
- Delete LegacyDestructive
Delete a dataset (legacy)
datasets.deleteLegacy
- ListRead-only
List datasets in a project
datasets.list
- List LegacyRead-only
List datasets (legacy)
datasets.listLegacy
- Remove DatapointDestructive
Remove a datapoint from a dataset
datasets.removeDatapoint
- UpdateDestructive
Update a dataset in a project; a datapoints list replaces its whole membership
datasets.update
- Update LegacyDestructive
Update a dataset (legacy); a datapoints list replaces its whole membership
datasets.updateLegacy
Metrics
- CreateWrite
Create a new evaluator metric
metrics.create
- Create VersionWrite
Create a new version snapshot for a metric
metrics.createVersion
- DeleteDestructive
Delete an evaluator metric
metrics.delete
- Deploy VersionWrite
Deploy a specific version of a metric
metrics.deployVersion
- ListRead-only
List all metrics and evaluators
metrics.list
- List VersionsRead-only
List versions of a metric
metrics.listVersions
- RunDestructive
Run an evaluator metric on an event
metrics.run
- UpdateWrite
Update an evaluator metric
metrics.update
Events
- CreateWrite
Create a new trace event
events.create
- Create BatchWrite
Create a batch of trace events
events.createBatch
- GetRead-only
Get a trace event by project and event ID
events.get
- Get LegacyRead-only
Get a trace event by event ID (legacy)
events.getLegacy
- SearchRead-only
Search trace events within a project
events.search
- Search LegacyRead-only
Search trace events (legacy)
events.searchLegacy
- UpdateWrite
Update a trace event
events.update
Configurations
- CreateWrite
Create a new configuration
configurations.create
- DeleteDestructive
Delete a configuration
configurations.delete
- ListRead-only
List LLM and pipeline configurations
configurations.list
- UpdateDestructive
Update an existing configuration
configurations.update
Projects
- CreateWrite
Create a new project in a workspace
projects.create
- DeleteDestructive
Delete (archive) a project
projects.delete
- GetRead-only
Get details for a project
projects.get
- UpdateWrite
Update a project name or description
projects.update
Sessions
- CreateWrite
Start a new session
sessions.create
- Create Event BatchWrite
Add a batch of events to a session
sessions.createEventBatch
Connect HoneyHive to Claude, ChatGPT or Cursor
- 1
Connect HoneyHive in Corsair
Add your HoneyHive API key once in Corsair. Corsair stores the credential encrypted and makes the HoneyHive API calls for your agent, so it never goes into a prompt or a config file. Connect HoneyHive
If your link lets the agent add apps, you can skip this: it gives you a sign-in link the first time it needs HoneyHive. On a ChatGPT personal plan, connect HoneyHive here first.
- 2
Add Corsair's MCP server to your agent
Every client uses the same address:
https://mcp.corsair.dev/mcpOpen Customize → Connectors → Add custom connector, paste the address, then select Connect.Add a custom MCP server with the address and choose OAuth. Personal plans get read tools only.Add to CursorOr add this to~/.cursor/mcp.json. Cursor opens Corsair to sign in the first time it connects.{ "mcpServers": { "corsair": { "url": "https://mcp.corsair.dev/mcp" } } }Run this, then run/mcpin Claude Code to sign in.claude mcp add --transport http corsair https://mcp.corsair.dev/mcpCodex, Gemini CLI, VS Code, Zed and other clients with OAuth sign-in work the same way. Clients without it, such as GitHub Copilot's coding agent, use a link's own address and an API key. Setup for every client
- 3
Sign in and choose what needs your approval
Your client opens Corsair to sign in. Pick a link: the apps your agent can use and a mode that decides what it may do without asking. In the default Cautious mode, HoneyHive's read and write operations run right away and its 11 destructive operations wait for your approval. Strict mode asks before every write and refuses destructive operations, and Read mode allows reads only.
FAQ
Install corsair and @corsair-dev/honeyhive, add the HoneyHive plugin, and connect with an API key. Then call methods like corsair.honeyhive.api.configurations.create() from your app. Corsair keeps each tenant isolated.
Yes. HoneyHive requires an API key from your HoneyHive account or developer dashboard. Create credentials with access that matches what you automate, then paste them into Corsair or connect through managed auth in Hub.
Yes. One Corsair client can hit HoneyHive, GitHub, Notion, Stripe, and the rest. Trigger HoneyHive from a webhook, push updates to Slack, or chain actions across plugins. Same tenant context across integrations, plain TypeScript.
Teams use Corsair for AI agent observability and evaluation platform for tracing, evaluating, monitoring, and improving LLM applications. Common builds include sync jobs, admin tooling, and agents that call HoneyHive APIs from your backend.
The SDK is open source. Self-host and you pay Corsair nothing for the integration layer. Hub bills for hosted auth and execution, not per HoneyHive API call. Hub has a generous free tier. For hobby projects you can probably run the whole thing on free. Paid plans start to matter with more tenants or heavier traffic.
We maintain @corsair-dev/honeyhive in the open repo. When HoneyHive ships API changes, upgrade the package. Your call sites stay typed. You should not need to rewrite the integration layer each time.
No. Corsair does not read your HoneyHive data or hold individual access credentials. Credentials stay in your database. API calls run on your instance. We do not store HoneyHive data on your behalf.
Yes. Add https://mcp.corsair.dev/mcp as a custom connector or MCP server in Claude, ChatGPT, Cursor, VS Code, Claude Code, Codex or any client that supports remote MCP servers with OAuth sign-in, then sign in to Corsair. On ChatGPT personal plans, custom servers currently get read tools only.
Yes. Put HoneyHive on a link in Strict mode and every write waits for your approval, while destructive actions are refused. The default Cautious mode asks only before the 11 destructive HoneyHive operations. You can change a link's mode at hub.corsair.dev/me.