Statsig for AI agents

Statsig is an experimentation and feature-flagging platform where product teams run A/B tests, roll out feature gates, and manage dynamic configs and metrics. 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 Statsig See the 93 actions ↓

Developer toolsOAuth 2.0 (MCP)MCP + RESTstatsig.com

AUDIT LOG · STATSIGPOLICY: acme-support

09:41:07 · claude · u_8f2read

statsig.search_tools

Search tools✓ allowed · 212ms

09:41:08 · claude · u_8f2read

statsig.get_context

Get context✓ allowed · 164ms

09:41:09 · claude · u_8f2write

statsig.create_gate

Create gate✓ approved · approved by user

09:41:12 · claude · u_8f2destructive

statsig.delete_experiment

Delete experiment✕ blocked · policy: deny

EVERY STATSIG CALL, ON THE RECORD

01 · USE CASES

What agents do in Statsig.

Create a new feature gate

Set up a feature gate for a rollout, then leave it unpublished until someone reviews the configuration.

Approve a pending experiment review

Approve a pending experiment review only after a teammate has signed off on the proposed test.

Check status of a gate review

Look up whether a feature gate review has been approved, cancelled, or is still waiting on sign-off.

02 · ACTIONS

93 Statsig actions, graded by risk.

Every Statsig 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.

statsig.get_context
Get context
statsig.search_tools
Search tools
statsig.get_audit_logs
Get audit logs
statsig.get_tool_schema
Get tool schema
statsig.get_gate_results
Get gate results
statsig.get_list_of_tags
Get list of tags
statsig.get_list_of_gates
Get list of gates
statsig.get_segment_by_id
Get segment by id
statsig.get_list_of_layers
Get list of layers
statsig.get_layer_overrides
Get layer overrides
statsig.get_list_of_metrics
Get list of metrics
statsig.get_list_of_prompts
Get list of prompts
statsig.query_logs_explorer
Query logs explorer
statsig.get_list_of_segments
Get list of segments
statsig.get_gate_review_by_id
Get gate review by id
statsig.get_layer_experiments
Get layer experiments

+ 30 MORE

write

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

statsig.create_gate
Create gate
statsig.create_layer
Create layer
statsig.create_prompt
Create prompt
statsig.update_prompt
Update prompt
statsig.create_segment
Create segment
statsig.update_segment
Update segment
statsig.create_autotune
Create autotune
statsig.create_experiment
Create experiment
statsig.create_gate_review
Create gate review
statsig.create_param_store
Create param store
statsig.update_param_store
Update param store
statsig.update_gate_entirely
Update gate entirely
statsig.create_dynamic_config
Create dynamic config
statsig.create_prompt_version
Create prompt version
statsig.update_layer_entirely
Update layer entirely
statsig.create_autotune_review
Create autotune review

+ 25 MORE

destructive

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

statsig.delete_experiment
Delete experiment
statsig.cancel_gate_review
Cancel gate review
statsig.delete_param_store
Delete param store
statsig.cancel_autotune_review
Cancel autotune review
statsig.cancel_experiment_review
Cancel experiment review
statsig.cancel_dynamic_config_review
Cancel dynamic config review

03 · HOW IT WORKS

Statsig in three steps.

01Your users connect StatsigThey sign in to Statsig on Arc0 Connect, under your brand, and approve the access you ask for.
02You set the rulesReads run, writes like “create gate” can wait for the user, and “delete experiment” can be denied outright.
03Any agent can actYour agent calls Statsig 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('statsig', {
  read: 'allow',
  write: 'ask',        // create_gate
  destructive: 'deny',  // delete_experiment
})

# 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 Statsig connects.

Statsig runs its own MCP server behind OAuth. Arc0 registers the client, users approve access on Arc0 Connect, and your agent reaches Statsig through the same endpoint, policies and audit log as every other app.

The same Statsig 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: OAuth 2.0 (MCP)
CREDENTIALS: Per-tenant encrypted vault
MODEL SEES: Results only, never credentials
AUDIT LOG: Every call, on every plan

05 · WORKS WITH

Use Statsig from any agent.

Claude

ChatGPT

Cursor

Codex

VS Code

OpenAI Agents SDK

Claude Agent SDK

Vercel AI SDK

Mastra

LangGraph

06 · RELATED

More developer tools apps.

Addressfinder17 actions

Addresszen2 actions

Alchemy36 actions

ALL DEVELOPER TOOLS APPS →

07 · FAQ

Statsig and Arc0, answered.

Q01

Can I use Statsig with Claude, ChatGPT or Cursor?

Yes. Connect Statsig 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 Statsig actions you allow.

Q02

How do users connect Statsig?

Q03

Which Statsig actions can my agent take?

93 in total: 46 read, 41 write and 6 destructive, such as “create gate”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Statsig?

Yes. Actions like “delete experiment” 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 Statsig too?

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

Get started

Plug Statsig into your agent.

Your users connect Statsig once, under your brand. Your agent gets 93 actions behind your policies, with every call on the record.

Connect Statsig Browse all apps →

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