Market and competitor research
Pythe searches the web, then writes a cited report on your market, audience, competitors, white space and risks. Choose how deep to go and ask your own questions.
Updated 2026-10-13
Research runs once during setup and any time you ask for it later, from HQ or with the run_research MCP tool. It is built for decisions, not for volume: relevance beats the number of competitors listed.
Choose the depth
| Depth | What you get | When to pick it |
|---|---|---|
| Quick scan | A fast read on demand and the obvious players. Fewer searches and a shorter write-up (about 6 to 12 competitors). | You want a first look and speed matters. |
| Standard | Competitors, audience, gap and risks. This is the recommended setting. | Most companies. |
| Deep dive | Everything in Standard, with more searches and more detail in the overview. | A crowded market, or a decision you cannot easily undo. |
You can also type questions you want answered. Three is usually plenty. They shape the research, and the report answers each directly with a confidence of high, medium or low.
How it runs
- Pythe validates your input.
- It runs a small, fixed batch of web searches. The number scales with the depth you chose.
- It makes one synthesis call to write the report.
- It checks the result so the app never sees a missing field.
- If the AI call fails, it returns a safe structured fallback marked as degraded, so the screen keeps working. A degraded result is labelled; it is never passed off as real research.
The whole run has to fit inside a fixed budget of 140 seconds, so every phase draws from it. You can close the tab while it runs. If the tab closes mid-run, Pythe picks the research back up when you return.
What is in the report
- Executive summary and market overview: long-form analysis, not bullets.
- Market size with TAM and SAM estimates, a growth outlook and key drivers.
- Trends and audience insights, including where your customers spend time and how they find tools.
- Competitors: evidence-backed, classified into categories, each with positioning, pricing where public, and a deep-dive you can open.
- Audience quotes with their sources.
- White space: where the market is under-served and why incumbents are unlikely to close it fast.
- Feature matrix across competitors.
- Market intelligence: maturity, demand signal, risks and threats, strategic insights, and quantitative data with 2 to 6 source citations.
- Your questions, each answered with a confidence.
After research
You review the results and can add notes, extra competitors and corrections. Those adjustments feed the plan build. Research is saved to your Library, and competitor deep-dives are saved as their own Library items.
Related
Previous: What Pythe can read: site, repo and files · Next: How the plan is built