Productivity · App · Browser extension
Recall AI knowledge base: one HN post, $120K and $9K MRR
Recall is an AI browser extension that builds a knowledge base from what you read. One Hacker News post saved the Recall AI knowledge base and led to $9K MRR.
| Revenue | About $9K MRRFounder-reported Source, 2023-10 |
|---|---|
| Team size | 3 full-time employeesFounder-reported Source, 2023-10 |
| Pricing | Not disclosed |
| Founded | Early 2022 (side project) |
| Founder | Paul Richards |
In November 2022, Paul Richards was ready to give up on Recall. The side project had a couple of thousand signups from Product Hunt and terrible retention. Before deciding, he posted on Hacker News asking honestly whether it was worth continuing. By bedtime the post was climbing; the next day it was number one, and within 24 hours an investor had wired $120,000 (founder-reported, Starter Story).
What problem does the Recall AI knowledge base solve?
People save articles, videos and bookmarks and almost never come back to them. Richards had the problem himself. As an early employee at Maltego, a company that maps relationships between pieces of information, he wondered whether the same idea could map connections between the things he read online.
Recall is an AI browser extension that stores content in a knowledge graph, categorises it automatically and finds links to things you saved before. When something related comes up, past content resurfaces, so saving actually leads to using.
Why did Recall almost fail after launch?
Richards started Recall as a side project in early 2022 while freelancing full-time. A first version after a couple of months let users save anything, such as people, books or films, as cards that were categorised and linked. He prepared carefully for Product Hunt, got a couple of thousand signups, and then watched almost nobody stay.
He parked the project. One persistent user kept asking for Dark Mode, so one weekend he built it, then posted the honest question on Hacker News. Close to 500 upvotes and about 13 hours at number one later, he had a funded startup (founder-reported, Starter Story).
How did watching users fix Recall’s retention?
Money did not fix retention. Richards hired a developer who soon became co-founder, which let him focus on users. Emailed surveys to churned users produced little. Sitting with people while they used Recall showed the truth: new users could not figure out how to use it.
“The biggest takeaway from this is that asking users to give feedback is far less valuable than just sitting down and watching them use the product.” — Paul Richards (founder-reported, Starter Story)
He also learned to charge earlier. He had kept postponing pricing at about 3,000 users, afraid of upsetting them.
What MRR has Recall reached?
By October 2023, Recall had about $9K MRR, growing 10% month over month, with three full-time employees (founder-reported, Starter Story). Total revenue and pricing are not disclosed.
| Channel | Result for Recall |
|---|---|
| Product Hunt (second launch) | Second product of the day; first ~20 paying users |
| AI tool directories | A steady source of paid users |
| Twitter bot | Monitored trending article tweets; worked briefly until Twitter’s API price rise |
| Content | Low-cost articles that attract users |
| Source | founder-reported, Starter Story |
How does Recall compare with Bluedot and Merlin?
Bluedot is another small AI browser extension that raised a modest angel round, and it also spent months on a product that did not sell before finding a self-serve format that did. Both founders learned that friendly feedback calls are not the same as real demand.
Merlin shows the scale an AI Chrome extension can reach, with a million-plus users. Merlin helps you do things on the page in front of you; Recall helps you remember what you read last month. That is a harder habit to build, which is why Recall’s onboarding problem mattered so much.
Which lessons from Recall apply to side projects?
- Ask the internet honestly. A candid Hacker News post brought users, attention and an investor.
- Watch users instead of surveying them. Sitting beside one confused user revealed more than forms.
- Charge earlier than feels comfortable. Postponing pricing at 3,000 users delayed revenue.
- Funding buys time, not product-market fit. Retention still had to be fixed after the money arrived.
Frequently asked questions
How much revenue does Recall make?
About $9K MRR, growing 10% month over month, as of October 2023 (founder-reported, Starter Story). Total revenue is not disclosed.
What does the Recall AI knowledge base do?
It lets you save content from the web, such as YouTube videos and articles, into a knowledge graph that categorises it automatically and resurfaces past content when something related comes up.
Is Recall funded?
Yes. After a Hacker News post reached number one in November 2022, an investor wired $120,000 on a SAFE within 24 hours (founder-reported, Starter Story).
How did Recall fix its retention?
By watching users try the product in person instead of relying on feedback forms, which showed that new users could not figure out how to use it.
Sources
- How My Passion Side Project Became A VC-Backed Startup In Just 24 Hours – Starter Story, 2023-10