AutoAGI is here.
Leading AI tech at your service.
Give it the big task. Get back the finished work. AutoAGI plans, researches and executes on frontier models, works inside 51 of your apps, and spins up private cloud harnesses when the work needs code.
From $29 a monthUsage in dollars, reset weekly51 connected appsCancel any time
AutoGPT Agent is now AutoAGI. Your account is ready.
Active AutoGPT Agent subscribers have been grandfathered into AutoAGI at their existing rate, while new memberships start at $29 a month. Sign in with the email on your subscription and your workspace opens on the Legacy plan.
Your member dashboard is right here.
Basic, Pro and Founder's Bundle: sign in with your checkout email for the setup guide, the Chrome extension, all three agents, every workflow, model transitions, the Workflow Generator and the AutoGPT agent.
Claude Anthropic
OpenAI GPT-5 series
Gemini Google
Grok xAI
Jev routing by TypeSafe
Codex OpenAI
Claude Code Anthropic
Hermes Nous Research
OpenCode open source
DeepSeekNot an answer. A deliverable.
Every task ends in something you can send, present or ship. Review it, ask for targeted revisions, export it, or turn it into a routine.
Research and writing, finished
Competitive briefs, positioning, long-form drafts and plans with sources, written to completion and exported to PDF, Word, HTML or Markdown.
Dashboards, calculators, presentations
Pro and Max tasks can return a working interactive result, privately hosted and shareable, instead of a static file.
Code and complex builds, delegated
Work that needs execution runs in a private cloud harness with its own files and session, then lands in your library with the outputs.
One brief in. One finished deliverable out.
AutoAGI treats every request as a job to complete, not a question to answer.
Give it the goal
Type it, dictate it, or drop in files and a website. Your business context and preferences travel with every task.
It plans and routes
Jev by TypeSafe scores the complexity. Economical models handle the simple parts. Frontier models take the hard ones.
It executes
Research, writing and analysis run autonomously. Code and complex builds delegate to a private cloud harness.
You get finished work
Review, revise, export. Alerts arrive the moment it is ready or needs your input.
Built for work that used to need a team
Autonomous research and writing
Bounded web research, structured plans and complete drafts with sources, run to completion without babysitting.
Private cloud harnesses
Code and file work runs in an agent workspace with its own session, metered on what the agent actually used within a compute ceiling. A run that hits its time limit keeps its workspace for the next one.
Jev analysis
Tag @Jev with a lead list, a set of pages or an export and state the criteria. Jev scores every item with calibrated probabilities and you get a report with a visual scorecard. On every membership.
Work inside your apps
Tag @Slack, @GitHub, @HubSpot or @Stripe in a brief and the task reads and writes where the work already lives. 51 apps plus webhooks, autonomous by default, read only when you say so.
Files, straight to your library
Ask for a Markdown brief, a CSV, a JSON export or an HTML page and it lands in your library as a downloadable file, no workspace needed. Code and binary builds still get a Cloud Harness.
Revisions and exports
Targeted revisions with version history, then PDF, Word, HTML or Markdown. Interactive results stay live.
Persistent context
Conversations, follow-ups and a private library of files and artifacts. Every task starts with what AutoAGI already knows about you.
Routines and next actions
Turn a task that worked into a routine on a schedule, and get next best actions suggested from your recent work.
Voice and alerts
Dictate a brief, listen to the result, and get in-app, email or browser push alerts when work lands.
51 apps. One tag in a brief.
Connect an app once. Then tag it in a task, a scheduled routine or a Cloud Harness run, and AutoAGI does the work inside it: reading, writing, filing and following up where your team already looks.
GitHub
Issues, pull requests, branches, file commits and releases. Push files to a repository from chat or a harness.
Slack
Read channels and threads, post messages, reply in threads and search your workspace.
Gmail
Search, read, draft, send and label email in Gmail.
Google Drive
Find, read, create and organize files and folders in Google Drive.
HubSpot
Contacts, companies, deals, notes and tasks in HubSpot.
Stripe
Customers, products, prices, subscriptions, invoices and payments in Stripe.
Zapier, n8n and anything with an HTTP endpoint connect through Webhooks. New apps are added as members ask for them.
Secure sign-in through Composio, our integration partner, and a policy you control. You approve permissions on each app's own page and your app password never enters AutoAGI. Every connection runs fully autonomous or read only, deletions and mass sends stay off until you turn them on, and you can scope it to the repositories or channels it may touch. Routines keep a snapshot of the connections they were saved with, Cloud Harness runs reach your apps through the AutoAGI gateway without ever holding a token, and every call is logged. Disconnect at any time.
Give ambition a place to become finished work.
Monthly billing. A dollar amount of usage every week, metered on what your work actually consumed and reset every 7 days. Every model and every Cloud Harness on every membership.
Founder
$6.70 of usage every week
Every model, every Cloud Harness and every connected app, from the first day. Turn a clear goal into polished, decision-ready work.
- Every frontier model, on your terms
- Claude Code, Codex and DeepSeek Cloud Harnesses
- Jev analysis: score leads, audit pages, review proposals
- 51 connected apps and webhooks, tagged from the composer
- Autonomous research, planning and finished deliverables
- PDF, Word, HTML and Markdown exports with targeted revisions
Pro
$23 of usage every week
Three and a half times the usage, with interactive results and Listen. Built for a week of real projects, not a handful of prompts.
- Everything in Founder, with 3.5 times the weekly usage
- Up to 6 tasks running at once and 10 active routines
- A higher compute ceiling on every Cloud Harness run
- Interactive dashboards, calculators and presentations
- Next best actions and Listen
Max
$46 of usage every week
Seven times the usage and the highest ceiling on every run. For teams of one who ship every day.
- Everything in Pro, with 7 times the weekly usage
- The highest compute ceiling per Cloud Harness run
- Up to 30 active routines and 75 scheduled runs a week
- Room for long multi-model builds every week
- Private hosted interactive results and reusable artifacts
Three invites a month, on every paid membership. Each one gives a friend $5 of usage to try AutoAGI free, no card needed, and your referral link earns you $5 of usage whenever someone joins through it.
Usage credits start at $10 inside AutoAGI for the heavy weeks. Autonomous tasks run until their deliverable, budget or execution limit and never consume resources after completion. Compare every detail and see the rate card.
Keep them. Add finished work on a weekly allowance you can see.
AutoAGI runs Claude Code, Codex and other agents in private cloud workspaces on its own weekly allowance, stated in dollars and metered on what your work actually consumed. Every task ends in a finished deliverable, and the allowance resets every 7 days at a fixed time for your account.
Every engine on every membership. Metered on what it used.
Claude Code, Codex, DeepSeek, Hermes and OpenCode are available from the first day. Every run has a stated compute ceiling that rises with your membership, a run that hits its time limit keeps its workspace for the next one, and usage credits start at $10 for the heavy weeks.
We started by automating ChatGPT. Now the whole job runs on its own.
AGI Layer shipped agent workflows and model transitions for ChatGPT Plus in 2023. AutoAGI is the same mission with the plumbing removed: no browser tab to watch, no prompt chain to maintain, and real cloud execution for the work that needs it.
| Product | What it is | Where you go |
|---|---|---|
| AutoAGI | Autonomous execution on frontier models and private cloud harnesses. Founder, Pro and Max memberships with weekly usage resets. | auto.agilayer.com |
| AutoGPT Agent | The original self-prompting agent. Active subscribers are grandfathered into the AutoAGI Legacy plan at their existing rate. | What changed |
| AGI Layer for ChatGPT | Prompt-chaining workflows in your own ChatGPT Plus tab. Still available, and every member gets everything. | Member dashboard |
Straight answers
What is AutoAGI?
How is AutoAGI different from a chat assistant?
What does a membership include?
I already pay for Claude Code or Codex. Why add AutoAGI?
What is Jev analysis?
Which models does AutoAGI use?
Which apps can AutoAGI connect to?
How do connected apps stay safe?
Is there a free trial?
I subscribe to the AutoGPT Agent. What happens to my subscription?
I bought AGI Layer before AutoAGI existed. Is my membership still honored?
Can I still buy the original AGI Layer product?
The State of Artificial General Intelligence
ChatGPT Automation in 2026: How to Run Hands-Free Workflows Without API Costs
ChatGPT automation in 2026 is shifting from prompts to hands-free workflows. Here’s why that matters now, and how AGI Layer helps users automate ChatGPT without added API costs.
The State of OpenAI in 2026: How Developers and Businesses Are Using ChatGPT, Codex, and GPT-Image-2
OpenAI in 2026 is shifting from chat to execution. Here’s how developers and businesses are using Codex, ChatGPT, and GPT-Image-2, and why AGI Layer makes hands-free automation practical without added API costs.
ChatGPT Automation Solutions: Zapier vs AGI Layer and the Best Options for Real Workflow Automation
Comparing the best ChatGPT automation solutions, including Zapier and AGI Layer, with a practical look at workflows, integrations, and where each tool wins.