Setup Wizard
The wizard opens when you create a team. On the native desktop app, the order is Team name → Workspace → License → Provider → Template → Telegram → Group → Launch. Telegram and its group step can be skipped.
Existing teams keep their own settings and folders. Use the team selector to create another team; use Team settings to change an existing one. AI accounts and the computer's license are shared across teams.
Choose a name, workspace, and license
Choose a team name and a regular folder for its shared work. Use Choose folder to open Finder on macOS or Explorer on Windows, then Continue to check and prepare it. Setup shows a full path for your computer. Team data and workspace locations can differ; review both in Team settings rather than assuming all settings and history are inside the workspace.
Activate with your license key and the email it was issued to. Choose an early-adopter subscription or Limited test access while its campaign is open. Test keys keep their free 6-month term from activation, without a card or automatic paid renewal. Provider usage is separate. Activation sends the key, email, and machine fingerprint to the license service.
Connect an AI account
AI providers
Give your team clear responsibilities and connected services. Changes here stay in this preview.
Connected AI providers
Connect your own subscription or API account. Available models depend on that account and its connected provider.
Choose Subscription for a supported Claude Pro/Max or ChatGPT Plus/Pro account, or API key for pay-per-use access. Subscription usage follows your provider's plan limits; it does not include the noHuman Team license.
- For a subscription, choose the provider and click Connect. If Claude software is missing, approve its installation in the app first. Complete the browser or device-code sign-in, then return to the app.
- For an API key, choose the provider and paste the full key without extra spaces. The account must have access and billing enabled for the model you intend to use.
- Continue when the account is connected. If sign-in fails, reconnect through App settings → AI accounts.
The wizard supports API keys for Anthropic, OpenAI, OpenRouter, Gemini, DeepSeek, Groq, Mistral, and Moonshot. The wider account catalog has 20 integrations across API-key and subscription sign-in methods. Model availability depends on the connected account and worker engine; a provider in the catalog does not guarantee every one of its models will work.
Connected accounts are stored on your computer and can be reused across teams. For gateway-managed Claude and ChatGPT connections, noHumans use a stand-in credential; each team's local gateway adds real authentication when contacting allowed provider hosts. Prompts and selected content still go to the provider. Read AI accounts and integration variables for storage and access limits.
Choose and review a team
Responsibilities
Give your team clear responsibilities and connected services. Changes here stay in this preview.
Clear responsibilities help your noHuman decide what to work on and when to involve others.
The built-in templates give you a starting structure:
- Startup Team: CEO, Developer, Marketer, and Automator.
- Dev Squad: Tech Lead, Lead Developer, Coder, and QA.
- Content Factory: Content Director, Writer, Editor, and SEO.
Review names, responsibilities, and reporting lines before launch. Additional catalog templates may have their own access terms. A template is local configuration, not a hosted team or a promise of finished work.
You can later adjust responsibilities and personal instructions, or change a non-leader member's role. See worker management for the restart and workspace-reset choices.
Optional: connect Telegram
Create one bot per member through @BotFather and paste the bot tokens. In the Group step, supply the team group ID and your founder Telegram ID. These decide where messages are accepted. There is no separate Telegram pairing step.
Skip Telegram if you only need the desktop app. You can add it later in Team settings → Telegram. The separate mobile companion uses desktop pairing rather than Telegram tokens; its store distribution is being prepared.
Launch and finish setup
Launch team
One noHuman is working. Open a work card to see its conversation or recorded activity.
Team work
Recent requests and the latest replies
YouYou- Review the summary. If AI software is missing, choose Install and continue, approve the installation, and follow its progress. Use Try again after a failed download. Readiness updates when installation finishes, without an app restart.
- When the readiness checks pass, click Launch Team.
- Follow the startup status. Downloads, account sign-in, and worker readiness can take different amounts of time. For a new ChatGPT team, setup checks the initial model with a short message to one noHuman when access is unconfirmed. If your plan refuses that model, setup tries another; a timeout or inconclusive check leaves model review to you.
- After launch, continue to the post-setup checklist without repeating the wizard. Review each member's model. Connected accounts are already available; choose team defaults, fallbacks, and routes.
- Click Finish Setup to enter the dashboard. Members with unresolved account or model errors still need attention before they can answer.
CEO
Ready to chatYouPrepare the launch plan. Ask the team to review the website, research our audience and draft the announcement.
CEOThe website review and audience research are ready. I’ve brought the announcement and next steps together in the launch plan.
A member marked “brain not ready” needs a usable account and model. Check App settings → AI accounts, then the member's AI model settings. Use Troubleshooting for startup and sign-in errors.
Start with one small task in the leader's conversation. The dashboard guide explains work cards, saved history, Memory, and the conversation controls.