Is CampaignGPT free?
Yes. The entire eight-section campaign is free with no email gate and no signup required. Anonymous users get 3 runs per tool per rolling 24 hours; a free account raises that to 10.
How long does a campaign take to generate?
Typically 2 to 3 minutes. The pipeline runs live web research first, then two large model calls: one for the strategy brief and one for all seven tactical sections. The report arrives complete rather than section by section.
Is the output actually researched, or just generated?
Both. Before generation, three live web searches run on your audience and pain point, your competitor landscape, and your campaign goal, collecting up to 15 deduplicated sources that ground the strategy stage. The sources are listed in your report. If research is unavailable, generation still completes from your inputs alone and the report shows no sources.
Where do the KPI benchmarks come from?
Benchmark ranges are generated for your specific channel mix and budget tier, then constrained to realistic industry caps enforced by the pipeline, for example email open rates between 15 and 45 percent and click-through rates between 1 and 5 percent. Treat them as calibration targets and replace them with your own baselines once you have two weeks of real data.
Why does my report look different from a colleague's for a similar product?
The strategy adapts to all six inputs together. A different budget tier changes the channel mix, a different sales cycle changes email spacing, and a different goal changes every CTA and KPI. Generative output also varies in wording between runs even when the strategic shape is the same.
Will two runs with the same inputs give the same campaign?
Within 24 hours, yes, exactly: identical inputs return the same stored report. After that window a new run generates a fresh campaign that follows the same strategic direction but with different wording and content ideas.
What happens if generation fails?
The report is marked failed and you can retry, which regenerates the whole campaign after a 30 second cooldown. Output is validated against strict schemas before being saved, so you never receive a structurally broken or half-complete report.
How do I get better output?
Sharpen the three text inputs. Name the exact buyer role instead of an industry, describe the pain in the customer's words, and make the product description specific enough that a stranger could tell you apart from competitors. The pipeline is instructed to reference your exact product and pain, so vague inputs are the main cause of generic output.