Chatbase: a 'ChatGPT for your PDFs' side project that reached $64K MRR in three months

Q: Why did Chatbase work?

First mover on an obvious need. "Chatbase was probably the first 'ChatGPT for your data' SaaS tool. This was a huge advantage since it wasn't that hard to make it go viral" (Indie Hackers). Full commitment once it moved. He "stopped doing anything university related" and failed two classes (Indie Hackers). Pricing for business use. "Most of the revenue is generated by the $399 monthly plan" (Indie Hackers). Channels after the hype. Product Hunt, story-style subreddit posts, Indie Hackers and AI directories (Indie Hackers).

Q: What can indie builders learn from Chatbase?

Speed matters most when a new capability lands. The MVP took him about two months, and he didn't do "any validation or market research" (Indie Hackers). That only works when the window is fresh. Viral launches don't repeat. He called influencer-driven virality "not a good long-term strategy" because influencers stop posting the same kind of tool (Indie Hackers). Move from a feature to a job. "Chat with a PDF" became a customer-support agent product (Chatbase), a use a business pays for every month. His rule on competition: "If there is no competition, you are either a genius or no one needs what you're building. If there is a lot of competition, it will be a race to the bottom" (Indie Hackers).

Q: What didn't work for Chatbase, and what are the risks?

Clones arrive fast. "Not sure how long this will last since now there are a crazy number of tools that are offering similar services" (Indie Hackers). Platform risk. "something can happen one day (e.g. a decision from OpenAI) that will change the trajectory of all these AI tools" (Indie Hackers). Low-tier churn. Most subscribers start on the lowest plan "mostly to test the chatbots", then upgrade "or churn" (Indie Hackers). All revenue numbers are self-reported in interviews; most of the 2025 article is behind a free Indie Hackers login.

Key takeaways

Self-reported MRR: $0 on 7 Feb 2023, $10,000 by 15 March, $64,000 by 13 May; most revenue came from a $399/month plan.
Being first to 'ChatGPT for your data' made it go viral, but the founder warned that clones and an OpenAI decision could change everything.
By February 2025 it reported $5M ARR as a B2B AI agent platform for customer support.

What it is

Chatbase started as "a 'ChatGPT for your PDFs' tool": upload a PDF and chat with its content (Indie Hackers, May 2023). Today it sells AI agents for customer support and sales "across chat, email, and voice" (Chatbase, checked 1 Nov 2026).

Timeline

2 Feb 2023: Yasser Elsaid, then in his final year of university, tweets the MVP to his 16 Twitter followers; it goes viral before it has a pricing page (Indie Hackers).
Feb–May 2023 (MRR, self-reported): $0 on 7 Feb, $400 on 11 Feb, $900 on 16 Feb, $3,000 on 28 Feb, $10,000 on 15 Mar, $64,000 on 13 May (Indie Hackers).
May 2023: hires two employees for features and support (Indie Hackers).
Feb 2025: relaunches on its two-year anniversary at $5M ARR as a B2B AI platform (Indie Hackers, Feb 2025).

Why it worked

First mover on an obvious need. "Chatbase was probably the first 'ChatGPT for your data' SaaS tool. This was a huge advantage since it wasn't that hard to make it go viral" (Indie Hackers).
Full commitment once it moved. He "stopped doing anything university related" and failed two classes (Indie Hackers).
Pricing for business use. "Most of the revenue is generated by the $399 monthly plan" (Indie Hackers).
Channels after the hype. Product Hunt, story-style subreddit posts, Indie Hackers and AI directories (Indie Hackers).

What to learn

Speed matters most when a new capability lands. The MVP took him about two months, and he didn't do "any validation or market research" (Indie Hackers). That only works when the window is fresh.
Viral launches don't repeat. He called influencer-driven virality "not a good long-term strategy" because influencers stop posting the same kind of tool (Indie Hackers).
Move from a feature to a job. "Chat with a PDF" became a customer-support agent product (Chatbase), a use a business pays for every month.
His rule on competition: "If there is no competition, you are either a genius or no one needs what you're building. If there is a lot of competition, it will be a race to the bottom" (Indie Hackers).

Risks and what didn't work

Clones arrive fast. "Not sure how long this will last since now there are a crazy number of tools that are offering similar services" (Indie Hackers).
Platform risk. "something can happen one day (e.g. a decision from OpenAI) that will change the trajectory of all these AI tools" (Indie Hackers).
Low-tier churn. Most subscribers start on the lowest plan "mostly to test the chatbots", then upgrade "or churn" (Indie Hackers).
All revenue numbers are self-reported in interviews; most of the 2025 article is behind a free Indie Hackers login.

FAQ

Short answers drawn from the sections above.

What is Chatbase?

Why did Chatbase work?

First mover on an obvious need. "Chatbase was probably the first 'ChatGPT for your data' SaaS tool. This was a huge advantage since it wasn't that hard to make it go viral" (Indie Hackers).
Full commitment once it moved. He "stopped doing anything university related" and failed two classes (Indie Hackers).
Pricing for business use. "Most of the revenue is generated by the $399 monthly plan" (Indie Hackers).
Channels after the hype. Product Hunt, story-style subreddit posts, Indie Hackers and AI directories (Indie Hackers).

What can indie builders learn from Chatbase?

Speed matters most when a new capability lands. The MVP took him about two months, and he didn't do "any validation or market research" (Indie Hackers). That only works when the window is fresh.
Viral launches don't repeat. He called influencer-driven virality "not a good long-term strategy" because influencers stop posting the same kind of tool (Indie Hackers).
Move from a feature to a job. "Chat with a PDF" became a customer-support agent product (Chatbase), a use a business pays for every month.
His rule on competition: "If there is no competition, you are either a genius or no one needs what you're building. If there is a lot of competition, it will be a race to the bottom" (Indie Hackers).

What didn't work for Chatbase, and what are the risks?

Clones arrive fast. "Not sure how long this will last since now there are a crazy number of tools that are offering similar services" (Indie Hackers).
Platform risk. "something can happen one day (e.g. a decision from OpenAI) that will change the trajectory of all these AI tools" (Indie Hackers).
Low-tier churn. Most subscribers start on the lowest plan "mostly to test the chatbots", then upgrade "or churn" (Indie Hackers).
All revenue numbers are self-reported in interviews; most of the 2025 article is behind a free Indie Hackers login.

Sources

Every link cited above, in order of first use.

1.Indie Hackers, May 2023 (indiehackers.com)
2.Chatbase, checked 1 Nov 2026 (chatbase.co)
3.Indie Hackers, Feb 2025 (indiehackers.com)