Use Gluecrawl in n8n
Build Gluecrawl scrapes into n8n workflows with the verified Gluecrawl node.
Gluecrawl ships an n8n integration, so you can create jobs, start runs, and read scraped rows from a workflow instead of writing the API calls yourself. The integration is built and maintained by Gluecrawl and verified by n8n, which means it is available on n8n Cloud as well as on self-hosted instances.
The package is @gluecrawl/n8n-nodes-gluecrawl (MIT, published by Deep Soup Labs). Its source is on GitHub.
Install the node
On n8n Cloud, open Settings → Community nodes → Install, enter @gluecrawl/n8n-nodes-gluecrawl, and confirm. Verified nodes are installable without any extra configuration.
On self-hosted n8n, use the same Settings → Community nodes → Install flow. Restart n8n if your deployment does not reload community packages automatically.
Once installed, search the node panel for Gluecrawl to add either node to a canvas.
Create the credential
The nodes authenticate with a Gluecrawl API key, the same key the REST API uses.
- Log in to your Gluecrawl dashboard
- Click your avatar → API Key → Generate API Key
- Copy the key — it is shown only once
In n8n, open a Gluecrawl node, choose Create new credential, select Gluecrawl API, and paste the key. Press Test to confirm it works before saving.
API keys are available on every paid plan. Free accounts can use the dashboard but cannot generate a key. See Authentication for the underlying scheme.
The Gluecrawl node
The Gluecrawl node exposes two resources, Job and Run. Each operation is a thin wrapper over one REST endpoint, so the reference pages below describe exactly what it sends and receives.
| Resource | Operation | Endpoint | Reference |
|---|---|---|---|
| Job | Create | POST /v1/jobs | Create a Job |
| Job | Get | GET /v1/jobs/{id} | Get a Job |
| Job | Get Many | GET /v1/jobs | List Jobs |
| Job | Delete | DELETE /v1/jobs/{id} | Delete a Job |
| Run | Start | POST /v1/jobs/{id}/runs | Create a Run |
| Run | Get | GET /v1/runs/{id} | Get a Run |
| Run | Get Many | GET /v1/jobs/{id}/runs | List Runs |
| Run | Get Items | GET /v1/runs/{id}/items | Get Items (JSON) |
| Run | Download CSV | GET /v1/runs/{id}/items/csv | Download Items (CSV) |
Job: Create takes a URL plus a goal or explicit columns, exactly like the API. See Goal or columns to decide which extraction input fits your target.
Run: Get Items emits one n8n item per scraped row, so the next node in the workflow receives records rather than a single blob. Run: Download CSV attaches the export to the item's binary property instead, for a file destination such as Drive or S3.
Billing is unchanged by running through n8n: creating a job costs 10 credits and is refunded if job processing fails, and runs are charged after completion based on protection level, listing pages, and detail-page visits. See Credits.
Wait for a job to be ready
A job must reach ready before a run can be created. This is the most common mistake in a first workflow: a Run: Start wired directly to Job: Create fires while the mapper is still working and fails with 409 job_not_ready.
Wire it one of these ways instead:
- Turn on Wait for Completion on Job: Create or Run: Start. The node polls in-process until the work finishes. It holds the execution — and on n8n Cloud, a concurrency slot — for the whole scrape, so keep it for short interactive workflows.
- Use the Gluecrawl Trigger to resume on the
Job ReadyorRun Completedevent. This is the right shape for anything scheduled or high volume.
A job that reaches failed is terminal. Gluecrawl never re-maps an existing job, so retrying against it keeps returning the same error — create a new job. See Jobs for the full status model.
The Gluecrawl Trigger node
The Gluecrawl Trigger node starts a workflow when Gluecrawl reaches a lifecycle state. It replaces the polling loop in the quickstart: the workflow stays idle while a scrape runs and wakes only when there is something to read.
| Event | Fires when |
|---|---|
| Run Completed | A run reached completed and its rows are available |
| Run Failed | A run stopped with an error; its credits are refunded automatically |
| Job Ready | Mapping finished and the job can be run |
| Job Failed | Mapping failed; the job is terminal and needs replacing |
Select one or more events on the node. An optional Job filter drops events for other jobs; leave it empty to receive events for every job on the account.
Activating the workflow registers a Gluecrawl webhook endpoint pointing at it, and deactivating removes the endpoint the node created. Deliveries are signed and verified before they reach your workflow. Delivery is at least once, so key any de-duplication on the payload's event_id. See Webhooks for the underlying delivery contract.
The usual event-driven shape is: Gluecrawl Trigger (Run Completed) → Run: Get Items → transform → destination.
Start from a template
Three ready-made workflows are published in the templates repository:
- AI newsletter — reruns one job per news site weekly, keeps every article seen so far in a data table, and asks an LLM to digest only what is new.
- Local business leads — mints a job for a fresh business type and city each week, then scores each business on how weak its web presence is.
- Property listing alerts — reruns one job per listing site each morning and emails only the homes that appeared since yesterday.
Download a template's .json, import it in n8n through Workflows → Import from File, then attach your Gluecrawl credential. Templates ship with credentials unset and arrive inactive, so review what a workflow does before you publish it — every run spends credits.