Use AI · no code needed

What would you like your AI to help with?

Start with a real team workflow. Each guide shows what you'll end up with, what you need first, and the exact prompts to give your connected assistant.

Not connected yet? Connect your assistant first — it takes a couple of minutes and one approval in your browser. Connect your AI →

Choose a workflow

Support · customer email

Run a customer support inbox

Triage support@, keep every request owned and on-status, prepare the routine replies.

Follow the support guide → Team email

Share an info@ or hello@ inbox

Everything gets an owner and is answered once — never twice, never not at all.

Follow the shared inbox guide → Sales

Qualify leads and keep the CRM fresh

Turn enquiries into contacts and deals, and rescue deals that go quiet.

Follow the sales guide → Client service

Keep client requests on track

Every request logged with a deadline, status answers on demand, painless handovers.

Follow the client service guide → Operations

Process orders and operational mail

Pull details into fields and flag the exceptions that need a person.

Follow the operations guide → Projects

Catch commitments made in email

Promises become cards, weeks get planned, Friday summaries get written for you.

Follow the project guide → Finance

Track invoices and chase payments

Amounts and due dates on every invoice, payment reminders drafted on time.

Follow the finance guide → Hiring

Manage a hiring pipeline

Applications summarised, stages kept current, scheduling replies prepared.

Follow the hiring guide → Personal

Delegate your own inbox

Triage on your rules, replies drafted in your voice, and a morning brief.

Follow the delegation guide →

Once a guide is working

Next step

Run it on a schedule

Hourly triage, morning digests and list watching, from Claude or ChatGPT scheduled tasks.

Look it up

Everything your AI can do

Every subject and action, each with a sample prompt.

Technical

MCP tool reference

Tool names, inputs and the permission each one needs.

Going further

Techniques and patterns for when the basics are running. You don't need any of this to follow a guide.

Prompts that work

The assistant sees Sortd through scope-gated tools (full vocabulary with prompts: AI reference), so a good prompt names the place (board and list) and the action — the model resolves ids itself via list_boards/list_tasks. Straight from live sessions:

Triage: "Check for new mail and file anything urgent onto the Support board's To do list; tag each task by topic."
Find & task: "Find the renewal thread from Jane and make it a task on Support / To do."
Prioritise: "Re-order my To do list by urgency — anything overdue on top."
Draft for review: "Draft replies to the three oldest unanswered threads on Support — leave the drafts in Gmail for me."
Report: "List everything still open past its due date across the Support board, grouped by assignee."
Board hygiene: "Move everything completed more than a week ago on AI Testing to the Closed list — don't change their completion state."

Three habits that make prompts reliable: name boards and lists by their visible names (the model matches them — ids are for code); say what must not happen ("do not complete them", "don't send, just draft") — the tools are separate, so the model can honour it precisely; and for anything destructive, ask it to list what it plans to change first.

Setups & permissions

Connect once per app — per-client steps for Claude, ChatGPT, Cursor and the rest. You can ask the assistant to set up its own connector; the browser consent stays yours.
Scopes are the job description. The scopes ticked on the consent screen (or on an API key) decide which tools the assistant even sees; send and delete scopes start unticked. Scope sets per job — triage, reply-prep, sending — are on the schedule page.
Tool permissions in the client: a sensible default is always-allow the read tools and keep per-use confirmation on writes — and always on sends. Sortd flags destructive tools (send, delete, secret rotation) so the client's permission UI can tell them apart.
Audit & revoke: every connected assistant (and every scheduled run) appears in the session list; revoke any one and it must reconnect.

Workflows

The triage loop — new mail → classify → task → label — is the connector's home turf; run it on a schedule (schedule workflows) and it becomes an hourly teammate.
Human-in-the-loop drafting: grant drafts.write but no send scope, and the assistant can prepare every reply while you keep the send button. Escalate to email.reply (thread-bound, structurally can't reach outsiders) only when a workflow has earned it.
Two-agent patterns: a scheduled agent files and tags; you (or another session) work the board interactively. Sortd's board is the shared state — no agent-to-agent plumbing needed, and every change is auditable and emits webhooks.

Agent runners & workflow platforms — push, don't poll

A chat assistant can't receive a webhook: it's a client with no public URL, which is why chat-side automation polls on a schedule. But agent runners and workflow platforms host endpoints — and then Sortd can push to them the moment something happens. The pattern is always the same three steps:

Get an endpoint URL from your platform (its webhook/HTTP trigger).

Subscribe it — needs webhooks.manage; the response shows the signing secret once:

curl -X POST "https://api.sortd.com/v2/webhooks/subscriptions" \
  -H "Authorization: Bearer sk_live_…" \
  -H "Content-Type: application/json" \
  -d '{"target_url":"https://YOUR-ENDPOINT",
       "events":["thread.reply_received","task.created","task.assigned"]}'

Verify the signature, then act — check x-sortd-signature (HMAC-SHA256, verification samples), then let the workflow call the Sortd API back with an API key, or hand the event to an agent step.

n8n

Add a Webhook trigger node (method POST) and copy its production URL — that's your target_url. Verify the signature in a Code node (Node.js crypto, constant-time compare), then branch: an HTTP Request node calls the Sortd API back (Authorization: Bearer sk_live_…), or an AI Agent node reasons over the event payload first — "a reply arrived on a tasked thread; decide whether to bump the task and draft a response".

Make (Integromat)

Create a Custom webhook trigger and use its URL as the target_url. Add a verification step over the raw body, then route into HTTP modules calling the Sortd API — or into Make's AI modules for classification/drafting decisions before the write-back.

Pipedream

An HTTP / Webhook trigger gives you the endpoint; each event runs your Node/Python steps with the raw body available for signature verification. Ideal when the "workflow" is really code — the same handler patterns as the webhooks guide, without hosting anything.

Zapier

Simplest of all: Sortd's Zapier app has native triggers — no endpoint, no subscription call, no signature handling. Pick the trigger, wire the action. Choose Zapier when its triggers cover your events; choose a webhook subscription when you need the full event list or custom filtering.

Self-hosted agents

Any service you run — an Express/Flask handler, a serverless function, a LangGraph-style agent service — can be the target_url. The webhooks guide has the exact verification code (Express and Flask); from there, trigger an agent run with the event as context, and let it call Sortd back through the API or its own MCP connection.

Choosing: assistants pull; servers push. An hourly triage is a perfect pull (schedule it); "react the moment a reply lands" is a push — use a runner from this page. Event list and delivery contract (signing, retries, replay, the delivery log): the webhooks guide.

Platform trigger names verified September 2026 — vendors rename UI periodically; the subscribe/verify/act pattern is unchanged when they do.