Internal research · Brevis
A study of 110 developer-facing AI products that grew from launch to wide adoption through earned, organic means: no paid spend, no borrowed fame. Built to inform how Brevis launches its AutoResearch framework and ARCA.
97% ship the product as the open-source repo itself. The GitHub page is the landing page and the star count is the public scoreboard; brand social accounts stay small by comparison.
Show HN, Launch HN, and the organic front page are where most of these went public. Roughly two-thirds touched HN at launch, more than any other external channel.
The most common breakout trigger is becoming a default inside a bigger tool (distribution-by-integration, 26%), ahead of any individual launch channel. For example: built into LangChain, Ollama, or the Hugging Face ecosystem, or going OpenAI-API-compatible.
45% led with one verifiable performance number, such as 2x faster with 70% less VRAM. This audience reruns claims, so a checkable result carries further than a product pitch.
Tech press reached 71% and peer founders amplified 38%. The same named voices recur (swyx, Simon Willison, Karpathy, Andrew Ng). Paid channels are near zero.