Keyword ideas from your seed topics
Expands the topics you already rank for into related search terms with their monthly volumes, appended to a Google Sheet.
Apps it connects
Authenticate each one once and the workflow is ready to run.
What it does
The integration steps this workflow runs, in the order it first runs them.
- 1
Google SheetsCreate tab - 2
Google SheetsAppend row - 3
AnthropicGenerate model response
How it works
Everything the template sets up, and what to fill in before the first run.
Once a month this takes the topics you already rank for, asks your keyword data endpoint what else people search around them, and appends every idea to a Google Sheet with the date it was collected.
It is a discovery workflow rather than a reporting one: it is not measuring the pages you already have, it is building the list of terms you have not written about yet. Each row records the date collected, the seed topic it came from, the keyword itself and its monthly search volume. Because every run appends and nothing is ever overwritten, the sheet doubles as a record of how each term's volume moves month to month, which is usually more telling than any single reading.
The columns are fixed and named by the workflow - Collected on, Seed topic, Keyword, Monthly searches - so it only ever writes figures it knows how to read. The first run creates the tab and writes that header; every run after that only appends.
The keyword data comes from an endpoint you supply, so it works with whichever keyword tool you already pay for. Fill in 'keywordApiUrl' with its address and 'keywordApiKey' with your key; the key is sent as a Bearer token in the Authorization header rather than in the address. Each seed topic is sent as one POST with a JSON body carrying "keyword", "limit" and "country", and the endpoint is expected to answer with a "keywords" list whose entries each carry a "keyword" string and a numeric "search_volume". An entry missing either is logged and skipped rather than written as a blank row, and a topic the endpoint cannot answer for is logged before the run moves on to the next one. Failure messages have the endpoint address taken back out of them before they are reported, so a key pasted into the address is never repeated back.
Once the sheet is written, Claude is asked which five terms are worth writing about next and why, and that shortlist is shown on the run. The step is optional: if Claude is not connected the ideas are still saved and the run still succeeds.
To set it up you need your seed topics, your keyword endpoint and its key, and an empty Google Sheet. 'sheetTab', 'maxIdeasPerTopic' and 'country' already have sensible defaults; the per-topic cap is enforced here as well as sent to the endpoint, so an endpoint that ignores it cannot flood the sheet.
Start from a workflow that already works.
Add "Keyword ideas from your seed topics" to your workspace, connect its apps, and make it yours. No credit card required.
Free plan available · No credit card required
