Grid82
grid82.comJoin the waitlist
BuildWorkflows overview

Automations overview

How Grid82 durable TypeScript workflows work, available triggers, built-in approvals and AI steps, and what a workflow cannot call.

A Grid82 workflow is one durable TypeScript file stored in an app’s Automations tab. Workflows run on durable serverless execution, meaning runs can sleep for months or pause for human approvals at zero compute cost while waiting.

Triggers

TriggerFires whenPayload inside trigger
schedule(expr, opts)A schedule or cron fires (e.g. "weekdays 09:00")firedAt
record.created("App.Table", { when? })A record is created (and matches when if provided)record
record.updated("App.Table", { fields?, when? })Specific fields change and match whenrecord, previous
record.deleted("App.Table")A record is removedprevious
manual({ input })Triggered manually by a user, API, or agentinput
webhook({ input })Signed POST to https://{suite}/hooks/{app}/{workflow}input

when clauses are evaluated before a workflow run starts, so non-matching changes consume zero workflow steps.

Capabilities and intentional boundaries

Workflows receive six SDK objects (trigger, db, ai, step, approval, http) and cannot use arbitrary npm packages, fetch, timers, or ambient clocks outside step.do:

  • Cross-app access. List every app the workflow touches in apps: ["Sales CRM", "Billing"]. An app workflow runs with Studio-level reach across the apps listed in apps, recording every write under the workflow’s own name.
  • Outbound HTTP (egress). List allowed domains in egress: ["hooks.slack.com"]. Secret URLs ("secret:SLACK_WEBHOOK") are configured write-only in suite Automations settings and injected securely by Grid82’s isolated outbound proxy.
  • No direct email or SMS primitive. Workflows cannot call notify or send emails directly. To alert someone, either send a http.post to a Slack/Zapier/Make/n8n webhook secret, insert a record into a monitored table, or call approval.ask to place a decision card in a role’s queue.
  • Human approvals. approval.ask(label, { role, question, show, timeout }) pauses the run until a member with that role approves or rejects in the Runtime (up to 30 days). Only fields placed inside show are exposed on the approval card.
  • AI decisions and generation. Call ai.yes, ai.choose, ai.score, ai.extract, or ai.write cleanly as durable steps.

Run lifecycle and testing

  • Check (workflows.check). Compiles TypeScript rules and renders the step diagram without saving a version.
  • Test (workflows.test). Runs real db queries and ai steps while capturing create, update, delete, and http.post side effects as simulated, skipping waits, and auto-approving approval.ask. Editor tests cost zero monthly workflow steps (capped per day by plan).
  • Activate. Saving compiles a version; activating switches live triggers onto that version. Runs already paused remain on the version they started on.

Use these docs with your AI tools

An AI agent can read this documentation directly. You do not need an account or an API key. Everything here is public and read-only.

Query these docs via MCP

Recommended

Add this server to Claude, Claude Code, Cursor, Mistral, or any tool that supports MCP. Your agent can then search Grid82 documentation and read it in full, instead of answering from memory.

https://grid82.com/docs/mcp
  • searchFind the passages that answer a question.
  • fetchRead one page in full, as Markdown.
  • list_pagesSee every page in this documentation.

Query these docs over HTTP

The same tools also work as plain web requests. Use this for scripts, or for any tool that does not support MCP. There is one endpoint per tool. Arguments go in the query string, and the answer comes back as JSON.

https://grid82.com/docs/api/docs/search?query=custom+domain

Read the OpenAPI description. It is built from the same definitions as the tools, so it always matches what the endpoints do.

Read these docs as Markdown

Add .md to any page URL to get its Markdown source. You can also send the headerAccept: text/markdown to the page URL itself.

To read the whole documentation in one file, open llms-full.txt. For a short index of every page, open llms.txt.