Tilda

Export and inventory Tilda sites with AI agents. Agents list projects and pages, read page HTML with its JS and CSS references, and export standalone documents with asset links rewritten for archiving, migration, or self-hosting.

7 actions API key available

Tilda website ↗ Platform Documentation ↗ Capability data verified 2026-07-27

A scheduled backup window opens, or a migration project kicks off

List Projects and Get Project establish what exists: sites, custom domains, export settings

List Pages inventories every page with its ID, alias, filename, and published date

The agent compares published dates against the last snapshot to find which pages actually changed

The site owner reviews the export plan, which pages, how many API calls it will cost against the hourly budget, and where the mirror will land, before the run starts

Export Page Full pulls each approved page as a standalone document with asset links rewritten, and every image downloads from its from URL to its to path in storage

A completion report posts with page counts, asset counts, and any pages that failed and need a retry window

What This Integration Enables

A Tilda site is a real asset: campaign pages, product stories, sometimes the entire web presence. Tilda's export API is how you make that asset yours, and this connector turns the export from an afternoon of clicking into an agent's routine. FlowRunner agents inventory projects and pages, read body HTML with its JS and CSS references for embedding or archival, and produce fully standalone documents with asset links rewritten to local paths for backup, migration, or self-hosting. The API's constraint is worth designing around rather than discovering: it is rate limited to 150 requests per hour, and a full export touches every page plus every asset reference. That finite budget is exactly why export runs in FlowRunner are planned, approved, and metered rather than fired blind, with a [human-in-the-loop](/concepts/human-in-the-loop) gate on the runs that spend it. - Back up entire Tilda sites to your own storage on a schedule, exporting only what changed - Migrate or self-host pages as standalone documents with all asset links rewritten - Snapshot page HTML plus JS and CSS references for embedding in other systems - Inventory projects and pages, IDs, aliases, published dates, to drive downstream automation Exported pages carry the "Made on Tilda" attribution required by Tilda's Terms of Service; workflows preserve it rather than strip it.

Without FlowRunner

The site exists only on Tilda Years of landing pages live solely in a builder account, one login mishap away from gone

Export is a manual chore Someone clicks through the export UI page by page, so backups happen before launches and never after

Migration is a rebuild Moving a page to self-hosting means recreating it, because nobody wants to chase asset URLs by hand

With FlowRunner

Snapshots on a schedule Agents inventory projects and export changed pages to your own storage without anyone remembering to

Assets resolved automatically Every image, script, and stylesheet arrives as a from/to pair, downloaded and filed at the path the HTML expects

Self-hosting is a workflow A page becomes a standalone document with rewritten links, ready to serve from your infrastructure

Use Case Scenarios

The backup that survives the account

Weekly, an agent walks List Projects, then List Pages per project, and diffs published dates against the previous snapshot. Only changed pages get Export Page Full. Each returned image downloads from its from URL and lands at its to path in [Amazon S3](/integrations/amazon-s3) alongside the HTML. The run report posts to [Slack](/integrations/slack) with what changed and what the run cost in API calls. If the builder account is ever locked, hijacked, or simply cancelled, the site exists somewhere the company controls.

Moving the marketing site to self-hosting

The team decides the campaign site should serve from their own infrastructure. The agent proposes an export plan from the page inventory, the owner approves it, and Export Page Full produces standalone documents while [SFTP](/integrations/sftp) places every asset at its rewritten path on the web server. Get Project supplies the recommended server config for hosting the exported files. The batch is paced across hourly rate-limit windows, and nothing about the pages changes except who serves them.

Landing page snapshots for the compliance file

A regulated business needs to prove what its pricing page said on a given date. On the first of each month, the agent calls Get Page Full for the designated pages and files each complete document, dated, into the archive bucket. When legal asks what the offer said in March, the answer is a file, not a memory. The same snapshots feed a diff report so marketing sees exactly what changed on pages they thought were static.

Human-in-Loop Highlight

Export Page Full at project scale is a spend decision against a hard budget: the Tilda API allows 150 requests per hour, and a sweep across a few hundred pages plus retries can consume every window for the rest of the day, starving the scheduled backup and anything else sharing the keys. So FlowRunner agents cost out the run first. The plan that reaches the site owner says what will be exported, why each page qualifies, how many calls it will take, and how the run will pace across rate windows. The owner trims or approves, and the same gate holds later at the moment that actually changes the business: an exported mirror does not replace the live site until a person has clicked through it. Inventory, diffing, and asset chasing run at machine speed; spending the budget and going live wait for a decision.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

7 actions

Projects

List Projects Lists all Tilda projects available to the connected account with ID, title, and description. Requires a Tilda Business plan with API access enabled. The starting point for inventory and backup sweeps.
Get Project Returns a single project's details: custom domain, export settings, favicon, 404 page ID, and shared images. Optionally returns the recommended web server config for hosting exported files.

Export

Export Page Exports a page's body HTML for self-hosting, returning the markup plus images, JavaScript, and CSS as from/to pairs: the original CDN URL and the local relative path the HTML now references. Each asset must be downloaded and stored at its to path before serving.
Export Page Full Exports a page as a complete standalone document with all asset links already rewritten to local relative paths, plus the images needed to serve it as from/to pairs. The workhorse of full-site backup and migration runs.

Frequently Asked Questions

What can FlowRunner do with Tilda?

FlowRunner agents can run List Projects, Get Project, and List Pages in Tilda, plus 4 more actions.

Does connecting Tilda to FlowRunner require OAuth?

No. Tilda connects to FlowRunner with an API key, no OAuth flow required.

Can Tilda trigger a FlowRunner workflow automatically?

Tilda doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Tilda

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us