Deep research with sources, written by the model you pick
Ask a question. Knecty searches the web, reads the pages it finds, and writes a report with sources, using the model you choose.
How a run works
- Quick overview: 1 web search, then the report.
- Standard: 3 web searches. The model splits your question into parts, writes the report, then checks it in a second pass.
- Comprehensive deep-dive: 6 web searches and more page text read, with the same check pass.
You choose the writer
Pick any chat model to write the report, the same list you chat with. The report shows its sources so you can check each claim.
Polite to the sites it reads
Pages are fetched by knectybot only when you start a run, and it honours each site’s robots.txt.
What it costs
- $0.01 per web search.
- The writing model’s actual token usage at the provider’s price, plus a 25% platform fee.
- You see an estimate and confirm it before the run starts.
- Searches that don’t end up running are not charged.
- If the model fails, you get a plain report built from the sources, and the model part of the charge is refunded. Searches that already ran stay charged.
Costs come out of the credit included with your plan, then any top-up. Every model’s price · Plans
Questions
Which model writes the report?
The one you pick before the run. Any chat model on your plan can write it.
How much does a run cost?
$0.01 per web search plus the writing model’s token usage with the 25% platform fee. Knecty shows the estimate first and you confirm it.
Can I follow up on a report?
Yes. You can ask a follow-up question on the report, and it is estimated the same way before it runs.
Read more: Deep Research mode · Help centre · Pricing
More features
Try it for 7 days
It’s part of the same subscription as every model and every other feature.