General Input for developers

General Input is built to be driven by agents, not just by people clicking through a dashboard. There are two supported ways in: the geni CLI, for anything running in a terminal or a CI job, and a Model Context Protocol server, for agents that speak MCP natively. Both give an agent the same thing: your connected accounts, your workflows, and your skills, with every credential use audit-logged.

Start building for free

Command line

The General Input CLI is published on npm as @general-input/cli and installs as geni:

npm install -g @general-input/cli
geni login

From there an agent can run workflows, manage credentials, search connected integrations, and run bash with your operator credentials injected as environment variables and scrubbed from the output. geni skills install teaches Claude Code, Codex CLI, Gemini CLI, and VS Code Copilot Chat how to drive it. The full command reference lives at docs.generalinput.com.

Model Context Protocol

Every workspace exposes an MCP server over Streamable HTTP, so Claude, ChatGPT, and any other MCP client can call your workspace as a set of native tools. The endpoint is workspace-scoped:

https://cloud.generalinput.com/v1/mcp/{organizationId}

Copy the exact URL from Settings → Connect AI in the dashboard. Authorization is OAuth 2.1 with PKCE and dynamic client registration, so a compliant client can register itself with no manual key exchange. Discovery documents:

Getting access

The free plan is $0 and includes 800 credits for the lifetime of the workspace. Sign up, connect an integration, and start building; there is no sales call in the way. Paid plans and per-credit pricing are on the pricing page.

For headless use, a CI job or a cron or an agent running without a browser, generate a workspace API key yourself under Settings → API Keys in the dashboard and hand it to the CLI. Keys are scoped to the membership that created them, can be revoked at any time, and every credential they touch is audit-logged with the reason given.

Machine-readable site

This site is built for crawlers and agents as well as people. Guides, comparisons, blog posts, integrations, prompts, and templates all have markdown twins: append .md to any page URL, or request the page with Accept: text/markdown.

  • llms.txt: indexed map of every section of this site
  • llms-full.txt: the full text of the site in one file
  • sitemap.xml: every URL, including all prompts, templates, and integrations

Missing pages return a real 404 with a markdown or JSON body pointing at those indexes, so an agent that guesses a URL wrong can recover without a human.