General InputQuickstart
Install geni, teach your AI agent about it, and start asking it to do things with your connected accounts.
geni is the General Input CLI. It lets an AI coding agent on your machine (Claude Code, Cursor, Codex, anything that can run shell commands) reach into the integrations you've connected (Slack, Gmail, GitHub, Stripe, your CRM, your calendar) and do things with them on your behalf, without ever seeing the actual passwords or tokens.
You don't need to write code. Tell the agent what you want; geni figures out which credential to use and how to call the right API.
1. Install
Install geni
# macOS / Linux
curl -fsSL https://generalinput.com/install.sh | sh
# Windows (PowerShell)
irm https://generalinput.com/install.ps1 | iexThe install script verifies you have bash, curl, and jq on your PATH and prints an exact brew install … / apt install … command if anything's missing.
Log in
geni loginOpens your browser to a one-time approval page. Approve, and you're authenticated.
Connect at least one credential
Open the credentials page and connect a service like Slack, Gmail, or GitHub. The credential lives in our cloud, encrypted; the agent only ever sees its id, not the secret.
You can also do this later by saying "I want to use Slack" to the agent. It'll prompt you to connect.
Teach your AI agent about geni
geni skills installDetects your AI coding agent (currently Claude Code, more coming) and drops a skill file the agent loads automatically. From now on, every conversation in that agent already knows how to use geni correctly. No copy-paste prompts required. Re-run after any geni upgrade to refresh.
Verify
geni doctorRuns through every layer (system tools, session, server reachable, skill installed) and prints a checklist. If anything's red, the line below it tells you exactly how to fix it. This is also the right command to paste into a support thread.
2. One-off questions
Talk to your agent like you'd talk to a colleague. Some examples to copy:
Send a message to my
#engineeringSlack channel saying "shipping the new feature today."
Look at my last 10 Gmail messages and summarize what I missed.
Open a GitHub issue on
acme-corp/websitetitled "Hero image is blurry on mobile."
What's my MRR this month? Pull from Stripe.
Add a calendar event tomorrow at 3pm called "design review" and invite jack@example.com.
The agent will:
- Run
geni credential listto find the right connected credential - Look up the right operation in the integration docs
- Run
geni exec bash --cred <id> --reason "..." -- 'curl ...'with your auth header injected - Show you the result
That's the whole loop. You describe the outcome, the agent figures out which integration + operation, and geni keeps your secrets safe — secrets never appear in the agent's transcript, every credential access is audit-logged with the reason the agent supplied.
3. Workflows
Coming soon
Saved, scheduled scripts that run in the cloud on cron, webhooks, or poll triggers, without your machine needing to be on. Same agent flow you use for one-off questions, but the agent saves the script as a reusable workflow that we run for you on a schedule.
A one-off question is something you ask the agent right now. A workflow is something you ask the agent to set up once and have us run forever. Every weekday at 9am, every time a Stripe payment fails, every time a new row lands in your spreadsheet.
Coming with workflows: trigger types (cron, webhook, poll), per-trigger config, per-workflow runtime logs, and the ability to enable, disable, or edit any saved workflow from the agent or the dashboard.
Where to go next
Browse what's connected
geni credential — list, get, connect new credentials.
Find operations to call
geni integration — search the integration catalog and read per-operation
reference docs.
Run things
geni exec bash — the execution primitive. One-line API calls, multi-step
pipelines, anything you'd put in a shell.