Aglyn REST API
The Aglyn REST API gives you programmatic access to your organization's data — datasets and records, contacts and the CRM around them — companies, pipelines, deals, tasks, activities, email templates and each site's leads — sites and their pages and what is on them, page templates, search listing, business details and form submissions, your store's orders and products, and your media library. Use it to sync content from another system, push orders into accounting, record shipments from a 3PL or warehouse system, keep a catalog and its stock in step with a PIM or warehouse, feed a catalog to a marketplace, sync a CRM's contacts and deals in either direction, log calls from a dialer, draft page changes for a person to review and publish, back up records, or build an integration.
The REST API is included on Starter, Pro and Custom. Create keys from Organization → Settings → API keys. What differs by plan is how much: Starter includes 500 requests a month and stops at 20,000, at 120 a minute for the whole workspace; Pro includes 1,500 with no monthly ceiling, at 600 a minute; Custom's limits are set by your agreement. See Rate limits & usage.
Base URL
https://app.aglyn.com/api/v1
All requests are over HTTPS and every response is JSON. The version lives in the path; there's no version header.
Quick start
Check your key and see what it can do:
curl https://app.aglyn.com/api/v1/me \
-H "Authorization: Bearer aglyn_sk_your_key_here"
{
"object": "api_key",
"org": "org_abc123",
"name": "Warehouse sync",
"scopes": ["datasets:read", "datasets:write"]
}
Then read some data — here, the first page of a dataset's records:
curl "https://app.aglyn.com/api/v1/datasets" \
-H "Authorization: Bearer aglyn_sk_your_key_here"
curl "https://app.aglyn.com/api/v1/datasets/ds_team/records?limit=100" \
-H "Authorization: Bearer aglyn_sk_your_key_here"
In JavaScript
const res = await fetch('https://app.aglyn.com/api/v1/datasets/ds_team/records', {
headers: { Authorization: `Bearer ${process.env.AGLYN_API_KEY}` },
})
if (!res.ok) {
const { error } = await res.json()
throw new Error(`${error.type}: ${error.message}`)
}
const { data, next_cursor, has_more } = await res.json()
Paging through everything
Lists are ordered by id and paged with an opaque cursor, so a full sync is a loop:
let cursor = null
const all = []
do {
const url = new URL('https://app.aglyn.com/api/v1/datasets/ds_team/records')
url.searchParams.set('limit', '100')
if (cursor) url.searchParams.set('cursor', cursor)
const page = await fetch(url, {
headers: { Authorization: `Bearer ${process.env.AGLYN_API_KEY}` },
}).then((r) => r.json())
all.push(...page.data)
cursor = page.next_cursor
} while (cursor)
Read ordering before you assume page 1 holds the newest records — it doesn't.
Machine-readable description
The whole API is described as OpenAPI 3.1 at:
GET https://app.aglyn.com/api/v1/openapi.json
It needs no key. A description of how to authenticate that itself required authentication would be useless at the only moment you want it, and it contains nothing this documentation does not already publish.
Point a generator at it and you get a typed client — every resource, every filter, the pagination envelope and the error shape, without hand-writing types this API already knows:
npx @openapitools/openapi-generator-cli generate \
-i https://app.aglyn.com/api/v1/openapi.json -g typescript-fetch -o ./aglyn-client
Two things to know when reading it:
- The schemas use JSON Schema 2020-12, so a field that can be
nullis a type union ("type": ["string", "null"]), not OpenAPI 3.0'snullable: true. A generator that only understands 3.0 will get optionality wrong. - Most write bodies close — an unknown field is rejected, not ignored. The
exceptions read only the members they document and ignore the rest:
datasets and records, whose record
valuesalso drop any field id the dataset's model doesn't define; creating a site; and media uploads.
MCP server
The same API is served as Model Context Protocol tools, for AI assistants, code editors and agents:
https://app.aglyn.com/api/mcp
Any MCP client can connect. Give it an API key as a bearer token, or give it nothing and it asks the person to sign in and pick a workspace and an access level — no key is copied anywhere. Either way, a tool call runs through the same pipeline a request does, with the same scopes, rate limit and usage meter.
The transport is Streamable HTTP and the server keeps no session. A client that is only given the address finds everything else on its own:
| Document | Address |
|---|---|
| Server card (SEP-2127) | https://app.aglyn.com/api/mcp/server-card |
Protected resource (RFC 9728) — named by the 401 | https://app.aglyn.com/.well-known/oauth-protected-resource/api/mcp |
| Authorization server (RFC 8414) | https://app.aglyn.com/.well-known/oauth-authorization-server |
| API catalog (RFC 9727), on every published site | /.well-known/api-catalog |
The tool list is the server's own answer to tools/list, generated from the
OpenAPI description, and depends on what the key or connection may do.
Service endpoints
Three endpoints need no scope; any valid key can call them.
GET /v1
What this API is and what it serves.
{
"object": "api",
"name": "Aglyn REST API",
"version": "v1",
"documentation": "https://docs.aglyn.com/api",
"openapi": "/api/v1/openapi.json",
"mcp": "/api/mcp",
"resources": ["sites", "media", "contacts", "companies", "pipelines", "deals", "tasks", "activities", "leads", "email-templates", "datasets"]
}
This lists only top-level resources. Form submissions, pages,
templates, orders
and products all live under a site
(/v1/sites/{siteId}/…) and are deliberately absent, so a client that walks this list
never builds a path that 404s. media is here because
/v1/media — the organization library — really is a top-level
path; each site's own files are additionally at /v1/sites/{siteId}/media. The five
CRM resources are organization-level like contacts, and so are top-level here. Leads are
a site's rows, but the path is /v1/leads with the site as a
parameter, so the path is what is listed.
GET /v1/me
Introspect the key you're calling with — useful for verifying a key after rotation, or
for failing fast at startup with a clear message. name is the name the key was given
when it was created, or null for a key from before names were required.
{
"object": "api_key",
"org": "org_abc123",
"name": "Warehouse sync",
"scopes": ["datasets:read", "datasets:write"]
}
Called by an app that signed in instead of using a
key, object is "connected_app" and name is the name the app registered under; the
other fields are the same. Its writes appear in activity as Connected app followed by
that name, where a key's appear as API key and the key's name.
GET /v1/usage
Where you stand against every plan band this month — requests, contacts, datasets and dataset storage — and, for each, whether crossing it bills or refuses; plus the size of each CRM collection, for sizing a sync. See Usage.
{
"object": "usage",
"month": "2026-08",
"apiRequests": { "used": 18422, "included": 100000, "remaining": 81578, "metered": true, "ceiling": null, "unitBytes": 262144 }
}
Resources
| Resource | Description |
|---|---|
| Datasets & records | Create, read, update, and delete datasets and the records inside them — the one resource the API can provision from nothing. |
| Contacts | Read your organization's contacts, add the people your own systems own, and edit their name, tags, notes and CRM profile. |
| Companies | The accounts your contacts work for — keyed by domain, with an owner and an address. |
| Pipelines | The stages a deal moves through. Read-only; seeded by the first deal. |
| Deals | Opportunities with a value, a stage and a status — create them, move them, close them. |
| Tasks | Calls, emails, meetings and to-dos with a due date, against a contact, company or deal. |
| Activities | What happened — a write-once log of calls, meetings and notes. |
| Leads | A site's work queue — everyone it has captured, with status, owner and notes — and the conversion of a lead into a contact, a company and a deal. |
| Email templates | The letters a team sends from a record — templates and snippets, with merge fields. |
| Sites | List sites, read their details, create a site, and publish one — including making chosen pages' drafts live. |
| Pages | A site's pages and folders — create, rename, re-address, move and delete them — what is on each page, edited as a draft until it is published, and each page's search listing. |
| Templates | The page templates a new page can start from: the site's own and the built-in starter pages. |
| Search listing | Read and change the title and description a site shows in search results, and how its page titles are built. |
| Business details | Read and change a site's business name, slogan, phone, email, address and social links. |
| Form submissions | Read a site's form submissions, mark them read as you process them, and delete them after export. |
| Orders | Read a site's store orders — line items, totals, refunds, disputes — and record shipments against them. |
| Products | Read and write a site's catalog — add, edit and delete products, and change variants' prices and stock. |
| Media | List and upload files in the organization library and in each site's media. |
Orders and products need a plan that includes commerce, in addition to their
scope — see each page's plan note. Companies, pipelines, deals, tasks, activities,
leads and email templates need a plan that includes the CRM suite (Starter and
above); a plan without it answers plan_required with code: "crm" on all seven, while
contacts keep working.
For event-driven integrations, see Webhooks — push instead of poll.
Everything you need to know
- Authentication — API keys and scopes.
- Rate limits & usage — 120 requests/minute per key, and how the monthly quota bills.
- Conventions — pagination, ordering, errors, and idempotency.