Jenna FREN Try it free
Do it yourself · With your AI agent

Hand Jenna to your AI agent

Everything in the app is also an API, and the API comes with a skill file written for agents. Give Claude Code, Codex, Cursor or any agent your API key and the skill, and it sets up the missions, runs them and reads the reports for you. Same product, same minutes, no extra cost.

Create your API key Read the skill file

Two minutes of yours, then the agent takes over · A key spends minutes, never money

What your agent can do, and cannot

With an API key
  • Create projects and missions: URL, objective in plain language, persona, mode, video format.
  • Launch runs, poll them to completion, read the verdict, the video, the UX report and the step log.
  • Fan a mission out over a list of values, one run per value, and re-run with a new list.
  • With the Autopilot subscription: store credentials as secrets, load reference material as knowledge, schedule routines.
  • Check the balance before a batch and stop when it is too low.
  • Publish a public share link for a project.
Only you, signed in to the app
  • Sign up and create or revoke API keys.
  • Buy packs, enable auto-refill, change the subscription.
  • Change the profile or delete the account.

A key gets a flat 401 on those routes. The skill file tells the agent to ask you rather than work around it.

Two minutes of yours

  1. Sign up

    In the app, with an e-mail address. 120 free minutes come with the account, no card needed.

  2. Create a key

    Settings, then API keys. Name it after the agent, for example "Claude Code". The key is shown once; put it in an environment variable or a secret manager, never in a repository.

  3. Decide the budget

    The agent can only spend what is on the account. Start with the free minutes or a small pack. Leave auto-refill off for an account an agent drives, unless you want it to keep going unattended. If the agent needs secrets, knowledge or a schedule, take the Autopilot subscription.

Connect your agent

The skill file at /api/jenna/skill.md is the complete contract: authentication, mission fields, the full flow with real calls, credits, errors, a worked example and a checklist. Any agent that can read a URL and make HTTP calls can use it.

Claude Code

The file is a SKILL.md. Install it once, export the key, then ask for what you want.

# install the skill for your user
mkdir -p ~/.claude/skills/jenna-api
curl -fsSL https://jenna.haredev.com/api/jenna/skill.md -o ~/.claude/skills/jenna-api/SKILL.md

# give it the key
export JENNA_API_KEY=jna_live_...

# then, in Claude Code
Use the jenna-api skill to audit the checkout of https://shop.example.com as a first-time visitor and tell me where it breaks.

Codex, Cursor and other coding agents

Point the agent at the file from its instructions file, so every session knows about it.

# AGENTS.md, .cursor/rules or the equivalent
For browser automations and UX audits, use Jenna. Read https://jenna.haredev.com/api/jenna/skill.md first
and follow it exactly. The API key is in $JENNA_API_KEY. Never try to buy credits or change billing.

Any other agent

No install at all. Paste this as the first message, with your key where the agent keeps secrets.

Prompt Read https://jenna.haredev.com/api/jenna/skill.md and follow it exactly. My Jenna API key is in the JENNA_API_KEY secret. Then: [what you want done]. Check the balance before starting, report the video and report links when the run finishes, and ask me if anything needs money.

What it looks like

You type one sentence. The agent does what a person would do in the app, then comes back with the evidence.

You

"Audit the sign-up flow of https://app.example.com as a novice, no narration, and give me the three worst friction points with timestamps."

The agent
  • Checks the balance.
  • Creates a project and a mission: that URL, persona Novice, mode walkthrough, no voice.
  • Polls the run until it finishes.
  • Reads the UX report and the step log, picks the three worst points, quotes the timestamps.
  • Returns the video and report links.
You, with Autopilot

"Set up a mission that logs in to our supplier portal with the credentials I will give you as secrets, reads the open orders page every morning at 7, and flags anything marked late."

The agent
  • Creates the mission with an objective that uses the built-in secret and page-reading actions.
  • Asks you for the credentials, stores them as mission secrets, never echoes them.
  • Runs it once to verify, then sets the routine.
  • Tells you what the first run saw and how many minutes it used.

Guard rails

Money

A key spends minutes, never money

There is no route an API key can call that charges your card. The worst an agent in a loop can do is use up the balance on the account, which is why you choose that balance.

Budget

Bound the balance yourself

There is no per-key spending cap yet. Keep auto-refill off on an account an agent drives, start with a small pack, and have the agent check the balance before batches, as the skill file instructs.

Control

Revoke in one click

Settings, API keys, revoke. The agent gets a 401 on its next call. Rate limits apply per key, and runs that stop making progress are ended by Jenna's own watchdogs.

Which path is yours

In the app

You type the objective

The simplest path for one site and a handful of journeys.

The automation page →
With your agent

This page

You already work with an agent, or you have many sites, many journeys or a pipeline to wire Jenna into.

With us

We build it and hand it over

Skilled or sensitive work judged against a standard, measured against your people before you rely on it.

The set-up service →

Questions developers ask

Can my agent spend money on my account?

No. An API key can spend the minutes already on the account and read the balance. Buying packs, enabling auto-refill, changing the subscription and managing keys all require a signed-in human session.

Does the agent path cost more?

No. Runs started by an agent are billed exactly like runs started in the app: the same per-minute price, the same packs, the same subscription.

Which agents work?

Any agent that can read a URL and make HTTP calls. Claude Code installs the skill file in one command; Codex, Cursor and others are pointed at the same file from their instructions file or a prompt.

Can the agent sign up for me?

Sign-up, key creation and anything involving money are done by you in the app. It takes about two minutes. Everything after that, the agent can do.

Is the skill file kept up to date?

Yes. It is served by the API itself and ships with every release, so it always matches the routes and fields that are live.

Your agent can run its first journey tonight

Create the key, hand over the skill file, type one sentence.

Create your API key Read the skill file