DocSeek
Turn team chats and docs into an AI knowledge base
spent today completely reworking docseek's pricing model. sharing because the before/after might be useful for other founders thinking about pricing.
BEFORE:
• starter at $5.99/mo with a 5-day trial (required credit card)
• pro at $12.99/mo
• enterprise tier with "contact sales" (lol, i'm a solo founder)
WHAT WAS WRONG:
• requiring a credit card upfront killed signups
• the enterprise tier was fake — nobody was contacting sales
• no free option meant people couldn't try it properly in 5 days
AFTER:
• free forever: 10 docs, 50 questions/month, no card needed
• starter at $7/seat: 50 docs, 300 questions/seat
• pro at $15/seat: unlimited everything
WHAT ELSE SHIPPED TODAY:
• document and question limits actually enforced now (before they were just display numbers)
• dashboard shows usage vs limits so users know where they stand
• when you hit the doc limit, we tell you upfront before you try to index — no silent failures
• cancelled subscriptions fall back to free instead of locking you out
BIGGEST LESSON: if you have a "contact sales" button and you're a solo founder with zero enterprise customers, just delete it. replace it with a generous free tier that lets people actually use your product.
site: https://docseek.co
curious — for those who've done pricing changes, did lowering the barrier (free tier) actually increase conversions to paid?
Hey IH!
I've been working on DocSeek and want to share what I'm building.
The problem
I noticed teams have the same issue everywhere — important knowledge gets buried in Telegram chats, Google Docs, and shared drives. Someone asks a question that was already answered last week. Nobody can find it. So they interrupt a coworker instead.
Multiply that by 10 people and you're losing hours every day.
What I built
DocSeek connects to your team's Telegram chats, Google Drive, and Google Docs. It indexes everything and lets anyone on the team ask questions through a Telegram bot. AI reads the context and gives an instant answer with sources.
Think of it as a team brain that actually remembers everything.
Where I'm at
- Live product, no paying customers yet
- Stack: Next.js + FastAPI + Claude AI + Qdrant
- Monetization: subscription per seat ($5.99–$12.99/user/mo)
What I learned so far
- Teams that use Telegram for work are massively underserved — most knowledge tools focus on Slack
- Nobody wants to build a wiki manually. Automatic indexing is the key
- The hardest part isn't the AI — it's getting the integrations right
What's next
Adding Notion and Slack integrations. Exploring a free tier.
Would love feedback
- Would you use something like this?
- What integrations matter most to you?
- How do you handle team knowledge today?
Happy to share more about the tech, pricing decisions, or anything else. AMA!
-
1
This is cool. The automatic indexing part is
I’ve seen teams struggle more with finding info than storing it.
How accurate are the answers when context is messy (like long chat threads)?-
1
thanks! that's exactly the problem we're solving — teams have the info, they just can't find it.
for messy telegram threads, the AI actually handles it pretty well. we chunk messages into context windows so it doesn't just match keywords — it understands the conversation flow. if someone asked "what did we decide about the pricing?" it'll pull the relevant part of the thread, not just any message that mentions "pricing."
that said, cleaner documents (google drive docs, PDFs) give the best answers. telegram threads are more like a bonus layer — the real value is when your team has a mix of both and can search across everything in one place.
the accuracy also improves as you add more docs — more context for the AI to work with.
want to give it a shot? free plan takes about 5 min to set up
-
Comment