Honeyhive for AI agents

HoneyHive is an AI observability and evaluation platform for testing and monitoring LLM application behavior. Product and engineering teams use it as a building block inside their own AI features. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

Connect Honeyhive See the 42 actions ↓

AI and modelsAPI keyMCP + RESThoneyhive.ai

AUDIT LOG · HONEYHIVEPOLICY: acme-support

09:41:07 · claude · u_8f2read

honeyhive.list_tools

List Tools✓ allowed · 212ms

09:41:08 · claude · u_8f2read

honeyhive.get_run

Get Evaluation Run Details✓ allowed · 164ms

09:41:09 · claude · u_8f2write

honeyhive.create_tool

Create Tool✓ approved · approved by user

09:41:12 · claude · u_8f2destructive

honeyhive.delete_dataset

Delete Dataset✕ blocked · policy: deny

EVERY HONEYHIVE CALL, ON THE RECORD

01 · USE CASES

What agents do in Honeyhive.

Pull events from a session

Fetch the events logged for a session to review how an AI feature responded to a user.

Compare experiment runs

Compare two evaluation runs to see which prompt version performed better.

Delete a dataset with confirmation

Delete an evaluation dataset only after a team member confirms it is no longer in use.

02 · ACTIONS

42 Honeyhive actions, graded by risk.

Every Honeyhive action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

Look things up. Allowed by default.

honeyhive.get_run
Get Evaluation Run Details
honeyhive.get_runs
Get Evaluation Runs
honeyhive.get_events
Get Events
honeyhive.list_tools
List Tools
honeyhive.get_metrics
Get Metrics
honeyhive.get_session
Get Session
honeyhive.get_datasets
Get Datasets
honeyhive.get_projects
Get Projects
honeyhive.get_run_metrics
Get Run Metrics
honeyhive.get_runs_schema
Get Runs Schema
honeyhive.get_events_chart
Get Events Chart
honeyhive.get_configurations
Get Configurations
honeyhive.get_events_by_session_id
Get Events By Session ID
honeyhive.compare_runs
Compare Experiment Runs
honeyhive.retrieve_events
Retrieve Events
honeyhive.retrieve_datapoint
Retrieve Datapoint

+ 3 MORE

write

Create and change things. Allow, or ask the user first.

honeyhive.create_tool
Create Tool
honeyhive.update_tool
Update Tool
honeyhive.create_event
Create Event
honeyhive.update_event
Update Event
honeyhive.create_metric
Create Metric
honeyhive.update_metric
Update Metric
honeyhive.create_dataset
Create Dataset
honeyhive.update_dataset
Update Dataset
honeyhive.update_project
Update Project
honeyhive.create_datapoint
Create Datapoint
honeyhive.update_datapoint
Update Datapoint
honeyhive.create_model_event
Create Model Event
honeyhive.create_configuration
Create Configuration
honeyhive.update_configuration
Update Configuration
honeyhive.create_batch_datapoints
Batch Create Datapoints
honeyhive.create_batch_tool_events
Create Batch Tool Events

+ 5 MORE

destructive

Delete, cancel or archive. Ask first, or deny outright.

honeyhive.delete_dataset
Delete Dataset
honeyhive.delete_datapoint
Delete Datapoint

03 · HOW IT WORKS

Honeyhive in three steps.

01Your users connect HoneyhiveThey add their Honeyhive api key on Arc0 Connect, under your brand. It goes straight into the vault.
02You set the rulesReads run, writes like “create Tool” can wait for the user, and “delete Dataset” can be denied outright.
03Any agent can actYour agent calls Honeyhive through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.

POLICY.TS

await arc0.policies.set('honeyhive', {
  read: 'allow',
  write: 'ask',        // create_tool
  destructive: 'deny',  // delete_dataset
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2

04 · AUTH AND DATA

How Honeyhive connects.

Users add their Honeyhive api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same Honeyhive connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH: API key
CREDENTIALS: Per-tenant encrypted vault
MODEL SEES: Results only, never credentials
AUDIT LOG: Every call, on every plan

05 · WORKS WITH

Use Honeyhive from any agent.

Claude

ChatGPT

Cursor

Codex

VS Code

OpenAI Agents SDK

Claude Agent SDK

Vercel AI SDK

Mastra

LangGraph

06 · RELATED

More ai and models apps.

Perplexity AI9 actions

All Images AI7 actions

AltText.ai10 actions

AltTextLab2 actions

ALL AI AND MODELS APPS →

07 · FAQ

Honeyhive and Arc0, answered.

Q01

Can I use Honeyhive with Claude, ChatGPT or Cursor?

Yes. Connect Honeyhive to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Honeyhive actions you allow.

Q02

How do users connect Honeyhive?

Users add their Honeyhive api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which Honeyhive actions can my agent take?

42 in total: 19 read, 21 write and 2 destructive, such as “create Tool”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Honeyhive?

Yes. Actions like “delete Dataset” are graded destructive. Set destructive actions to deny, or to ask so the user approves each one, and blocked calls still show up on the audit log.

Q05

Can my own backend call Honeyhive too?

Yes. The same Honeyhive connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Honeyhive into your agent.

Your users connect Honeyhive once, under your brand. Your agent gets 42 actions behind your policies, with every call on the record.

Connect Honeyhive Browse all apps →

Free to build · MCP + REST · Audit log on every plan