Scheduled: ongoing inputs for your Parser
Scheduled is a feature of Colbot Parsers that fetches source content at regular intervals. It connects sources you want to monitor to a Parser and submits changed content for AI extraction. The Parser uses its existing fields, instructions, and review workflow to organize the results in Google Sheets.
A Parser can have multiple scheduled inputs, each with its own source, interval, and additional extraction instructions. For information you need to collect repeatedly, Scheduled reduces the work of checking sources, copying content, and submitting it for extraction.
From ongoing collection to structured results
Scheduled fetches source content at regular intervals. The Parser interprets that content and extracts the fields you need. Together, they provide an ongoing workflow from source content to Google Sheets:
- Fetch source content. Content is fetched automatically in the cloud using each input’s source settings and interval. Runs do not depend on your computer or browser staying open. Cloud fetching does not use your local browser’s login session, so the source must meet the relevant access requirements.
- Check for changes. The fetched input is compared with the last input successfully submitted to the Parser. Identical content does not create a new parsing task.
- Extract the fields. Changed content is sent to the Parser, which produces structured results using the sheet’s columns and extraction instructions.
- Review and write. Results follow the Parser’s review settings before being written to Google Sheets.
Change detection compares the input submitted to the Parser; it does not mean that only new records will be extracted. The fields and records extracted depend on the Parser’s settings. See billing when content changes little or not at all.
Supported input sources
The input source determines where Scheduled fetches content and which fetching options are available. It does not change the Parser’s target fields or review workflow. The current version supports the following source types.
Web page
Web page sources fetch public pages, such as product announcements, release notes, event listings, and product information. For example, a Parser can organize announcements into Google Sheets columns for title, publication date, summary, and link.
For complex or dynamically loaded pages, you can select page areas to include or exclude and prepare the content with waiting, clicking, or scrolling options. Fetching results depend on the target website’s page structure and access conditions.
- Setup and example: See the scheduled announcement collection example for the complete workflow, from setting the source and interval to viewing extraction results.
- Web capture comparison: Compare Scheduled and Web Clipper by scheduled runs, login sessions, content selection, and page interactions.
JSON API
JSON API sources fetch data from an endpoint for the Parser to extract and organize. The response does not need to match your Google Sheet’s field names or structure. The Parser can use your instructions to select information from nested data and map it to the target columns.
GET and read-only POST requests are supported, with request headers and a JSON body as needed. Use API sources for release records, catalog information, and other regularly updated data available through an endpoint.
Shared Scheduled settings
Every scheduled input has its own interval and additional extraction instructions, regardless of its source type. These determine how often content is fetched and how the Parser should process that source.
Interval
The interval starts when the current fetch run finishes. It does not wait for the Parser to finish parsing the content submitted after fetching.
For example, with a one-hour interval, a fetch that finishes at 10:05 is followed by the next automatic run at 11:05 or later, rather than on the hour.
Automatic runs may start slightly later than the selected interval. This is an interval setting, not a guarantee of execution at an exact clock time.
Additional extraction instructions
Additional extraction instructions are optional and apply to the Parser after fetching. They do not control how the source is fetched. They are included with every task created by this scheduled input to clarify which information to extract or how to process it.
For example, a release feed might use “Extract only stable releases and exclude prereleases.” Multiple sources connected to the same Parser can have different instructions while still using that Parser’s target fields.
Changing these instructions alone does not force unchanged source content to be parsed again. A new parsing task still depends on whether the fetched content has changed.
Billing and usage
Requirements for running Scheduled
The Parser’s team must have an active Basic subscription to run Scheduled.
Scheduled usage is billed to the Parser’s team, with fetching and parsing charges recorded separately in Credit History.
What the charges include
A complete Scheduled workflow has two charges: fetching and Parser task parsing.
| Charge | What it covers |
|---|---|
| Fetching | Retrieving content from the source, billed according to that source’s charging rules. |
| Parser task parsing | AI parsing performed by the Parser task created after fetching, billed separately under the normal task rules. |
Prices shown when choosing a source cover fetching only. They do not include Parser task parsing.
Billing and Credits FAQ:
- How are web page fetches charged?
- How are JSON API requests counted?
- Can a failed run still incur charges?
- How am I charged when input content changes little or not at all?
FAQ and detailed rules
See the full Scheduled FAQ, or browse other topics in the FAQ directory.