Runs
A run is one full cycle of the Wonderblogs content pipeline: from researching a topic to delivering a finished blog post. Think of it as pressing "go" on your content engine.
What happens during a run
Each run follows a structured pipeline to produce high-quality, on-brand content:
- Context fetch: Wonderblogs calls your Service URL to learn about your brand, audience, and content preferences.
- News scan: When your content orientation is set to Balanced or higher (3–5), Wonderblogs searches the web for current news and developments related to your topic. This grounds the next step in real, timely information. The depth of the scan increases with your orientation setting. If the scan fails, the run continues without it.
- Topic refinement: AI analyzes your content focus, existing articles, and the news scan results to find a unique angle that avoids duplicating what you've already published.
- Deep research: A focused deep dive on the refined angle gathers supporting facts, statistics, and expert insights.
- Writing: An AI model drafts a complete blog post with title, content, SEO metadata, and tags.
- Evaluation: First, a deterministic check validates formatting rules (dashes, banned phrases, required components). Then, a separate AI review checks the draft for quality, relevance, and brand alignment. The post may be revised based on feedback.
- Cover image: A cover image is generated as a background task while the post content is finalized.
- Publishing: Once the image is ready (or if it fails), the finished post is delivered to your Webhook URL.
Source file uploads
You can attach a source document in two places: the Run Now modal in the dashboard, or the POST /v1/posts/generate-from-file API endpoint.
Wonderblogs extracts text from the file and uses it as primary source material for the post.
Supported formats:
- PDF (text extraction is best-effort; image-based PDFs may not work)
- TXT, Markdown (.md)
- CSV (column headers + first 100 data rows are used)
- JSON, XML
The file is parsed into text and the original binary upload is discarded. If extracted text is longer than 20,000 characters (about 5,000 words), Wonderblogs automatically summarizes it before writing. Otherwise, the extracted text is used directly.
Important: When attaching a source file, you must also provide a topic. The topic tells the AI what angle to take on your source material. Without a topic, the run will be rejected.
File size limits depend on your plan: 500 KB on Free, 2 MB on Pro.
On the run detail page, runs with an uploaded file show a Source File card with the filename. If summarization was used, the card includes a SUMMARIZED badge.
During generation, source content is wrapped in a <source_document> block and treated as factual reference data. Instructions inside that document are ignored, and the evaluator checks that the draft uses the source content appropriately.
Run modes
Wonderblogs supports three run modes:
- Create: The default. Generates a brand-new blog post from scratch, starting with topic selection and research.
- Rewrite: Takes an existing blog post and rewrites it using the same quality pipeline. The original content is used as a starting point for the AI writer, and the rewritten post is linked back to the original. See the Rewrites help page for details.
- Translate: Translates existing content into a target language. You can start a translation from the Run Now modal (select the Translate tab, paste content, and pick a target language) or via the API endpoint
POST /api/v1/posts/translate. The translated post is delivered via the same publish webhook or callback as other runs.
Three ways to trigger a run
- Manual: Click Run Now on the Runs page to start immediately. You can create a new post (optionally with a topic or source file), rewrite an existing one by pasting its content, or translate content into another language.
- Scheduled (cron): Set a cron schedule in your site settings and Wonderblogs handles the rest automatically.
- API: Trigger runs programmatically from your own systems using a per-site API key. See the API Documentation for details.
Run statuses
| Status | What it means |
|---|---|
| Executing | The run is currently in progress. |
| Completed | All steps finished successfully and the post was delivered. |
| Failed | The run encountered an error. Check the log for details. In some cases, Wonderblogs automatically retries the run with an alternative configuration. If the retry succeeds, you'll see the finished post in your normal post list. |
| Cancelled | The run was stopped by an administrator before it could finish. |
Run logs
Every run records a detailed step-by-step log. Click on any run to see what happened at each stage, with timestamps and feedback from the AI review. This makes it easy to understand the result and diagnose any issues.
Live progress
While a run is executing, the run detail page updates in real time. You can watch the workflow progress live as it moves through research, writing, evaluation, and publishing.