Get started

Lakera Guard for AI agents

Lakera Guard is a security layer for AI applications that screens prompts and content for injection attempts and safety risks. 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 Lakera GuardSee the 10 actions ↓
AI and modelsAPI keyMCP + RESTlakera.ai
AUDIT LOG · LAKERA GUARDPOLICY: acme-support
09:41:07 · claude · u_8f2read
lakera_guard.get_policy
Get Policy✓ allowed · 212ms
09:41:08 · claude · u_8f2read
lakera_guard.get_project
Get Project✓ allowed · 164ms
09:41:09 · claude · u_8f2write
lakera_guard.create_policy
Create Policy✓ approved · approved by user
09:41:12 · claude · u_8f2destructive
lakera_guard.delete_policy
Delete Policy✕ blocked · policy: deny
EVERY LAKERA GUARD CALL, ON THE RECORD
01 · USE CASES

What agents do in Lakera Guard.

01

Screen content for injection risk

Screen an incoming prompt or document against a guardrail policy before it reaches a model.

02

Check a project's active policy

Look up a project's current guardrail policy to confirm what it blocks.

03

Approve before deleting a policy

Require sign-off before deleting a guardrail policy, since it removes active protection.

02 · ACTIONS

10 Lakera Guard actions, graded by risk.

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

read

2

Look things up. Allowed by default.

  • lakera_guard.get_policy
    Get Policy
  • lakera_guard.get_project
    Get Project

write

6

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

  • lakera_guard.create_policy
    Create Policy
  • lakera_guard.update_policy
    Update Policy
  • lakera_guard.create_project
    Create Project
  • lakera_guard.update_project
    Update Project
  • lakera_guard.screen_content
    Screen Content
  • lakera_guard.evaluate_detectors
    Evaluate Detectors

destructive

2

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

  • lakera_guard.delete_policy
    Delete Policy
  • lakera_guard.delete_project
    Delete Project
03 · HOW IT WORKS

Lakera Guard in three steps.

  1. 01Your users connect Lakera GuardThey add their Lakera Guard api key on Arc0 Connect, under your brand. It goes straight into the vault.
  2. 02You set the rulesReads run, writes like “create Policy” can wait for the user, and “delete Policy” can be denied outright.
  3. 03Any agent can actYour agent calls Lakera Guard 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('lakera_guard', {
  read: 'allow',
  write: 'ask',        // create_policy
  destructive: 'deny',  // delete_policy
})

# 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 Lakera Guard connects.

Users add their Lakera Guard 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 Lakera Guard 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 Lakera Guard 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 actionsAffinda119 actionsAgent Mail5 actionsAI/ML API30 actionsAivoov1 actionsAll Images AI7 actionsAltText.ai10 actionsAltTextLab2 actions
ALL AI AND MODELS APPS →
07 · FAQ

Lakera Guard and Arc0, answered.

Q01

Can I use Lakera Guard with Claude, ChatGPT or Cursor?

Yes. Connect Lakera Guard 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 Lakera Guard actions you allow.

Q02

How do users connect Lakera Guard?

Users add their Lakera Guard 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 Lakera Guard actions can my agent take?

10 in total: 2 read, 6 write and 2 destructive, such as “create Policy”. Your policies decide which of them each agent may call.

Q04

Can I stop my agent from deleting things in Lakera Guard?

Yes. Actions like “delete Policy” 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 Lakera Guard too?

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

Get started

Plug Lakera Guard into your agent.

Your users connect Lakera Guard once, under your brand. Your agent gets 10 actions behind your policies, with every call on the record.

Connect Lakera GuardBrowse all apps →

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