An AI teammate for Lakera Guard.
Connect Lakera Guard and your Cicada agent works in it for you: 10 Lakera Guard actions, from looking things up to making changes. Ask from iMessage, WhatsApp, Slack or the app, and keep it read-only if you only want it to look.
300 free credits. No card. Read-only if you want.
- What broke in Lakera Guard overnight, and is it new?
- Open a Lakera Guard issue for this bug with the steps from the Slack thread.
- Summarise this week's Lakera Guard activity for the standup.
- Tell me when something in Lakera Guard needs my attention.
10 Lakera Guard actions, picked for what you ask.
About Lakera Guard: Lakera Guard screens AI application interactions for security and safety risks and provides Enterprise APIs for managing guardrail policies and projects.
Reads
4 in all · always allowed- Get PolicyRetrieve an Enterprise Lakera policy by ID, including its detector configuration, custom…
- Get ProjectRetrieve an Enterprise Lakera project by ID, including its assigned policy, metadata, and…
- Screen ContentScreen the latest LLM or agent interaction with Lakera Guard and return the policy decision.
- Evaluate DetectorsEvaluate content and return one confidence result per configured detector without producing an…
Changes
6 in all · off when Lakera Guard is read-only- Create PolicyCreate an Enterprise Lakera policy with built-in detectors and/or custom guardrails.
- Delete PolicyPermanently delete an Enterprise Lakera policy by ID. The provider rejects deletion while any…
- Update PolicyReplace an Enterprise Lakera policy's complete configuration and create a new policy version.
- Create ProjectCreate an Enterprise Lakera project, optionally assigning a policy, metadata, and detect-or-flag…
- Delete ProjectPermanently and irreversibly delete an Enterprise Lakera project by ID.
- Update ProjectUpdate an Enterprise Lakera project by ID. Name is required by the official request schema;
Showing 10 of 10. Your agent can use every one of them; you don't pick them, you ask for the outcome.
Connected in a minute. In your control after that.
Connect Lakera Guard
Open Connections, find Lakera Guard and paste an API key from your Lakera Guard account settings. Cicada checks it works before saving it.
Choose what it may do
Full, Ask first or Read-only, for the whole workspace. Share Lakera Guard with the team, or keep it under each person's own login.
Ask from anywhere
Text your agent on iMessage or WhatsApp, message it in Slack or the app, or put a Lakera Guard job on a schedule: every morning, every Friday.
Lakera Guard, together with the rest of your stack.
Ask for an outcome that spans apps and your agent pulls from each one, cross-checks them and sends back one answer.
More developer tools integrations
Often used alongside
Lakera Guard and Cicada, answered.
Does Cicada integrate with Lakera Guard?
Yes. Lakera Guard is one of the 888 apps with a page here, and it connects with an API key. Once it's connected, every agent in your workspace can use it from any channel you talk to them on.
How do I connect Lakera Guard?
Open Connections, find Lakera Guard and paste an API key from your Lakera Guard account settings. Cicada checks it works before saving it.
What can Cicada do in Lakera Guard?
10 actions: 4 that read from Lakera Guard and 6 that change things in it, for example get policy, get project, screen content. It picks the right ones for what you ask and chains them with your other apps.
Can I stop Cicada changing things in Lakera Guard?
Yes. In Connections, set Lakera Guard to Read-only and no agent in your workspace can create, change or delete anything in it; it can still read. Ask first lets it act but makes it check with you before each change.
Can I ask about Lakera Guard from my phone?
Yes. Text your agent on iMessage or WhatsApp, or message it in Slack or the app. It answers in the same thread, and it can run Lakera Guard jobs on a schedule, like every morning at 8.
Is my Lakera Guard data safe?
Your Lakera Guard login or key is stored on the server and never appears in a conversation. Disconnecting removes it. Every action an agent takes is in your activity log, and your data isn't used to train models.
One teammate. The output of a team.
Works nights, remembers every decision, answers in the thread. Start free with 300 credits.