LightfieldAI-native CRM that builds itself and does work for you
Promoted
ZeeDocs
AI chatbot that answers from your own documents & manuals
6 followers
AI chatbot that answers from your own documents & manuals
6 followers
ZeeDocs lets businesses and individuals turn their PDFs, Word docs, and manuals into a smart chatbot that answers questions instantly - and even shows the right diagrams and images from your files. No coding required.
Hey Product Hunt! š
I'm, a long-time IT Professional and first-time SaaS founder. Excited to share ZeeDocs, a document-reading chatbot I've been quietly building.
The problem I kept seeing in client work:
Every team had the same pattern. Critical information lived in documents nobody read. Sales reps re-asked product questions. Support staff thumbed through 200-page manuals. New hires kept asking HR the same policy questions for the tenth time. Documentation existed; people just couldn't find answers inside it.
What ZeeDocs does:
Upload your documents (PDF, Word, text). Embed a chat widget on your website or use it directly from the dashboard. Visitors and team members ask questions in plain language. The bot answers grounded only in your content, with the source document and page number on every response.
What I focused on while building:
ā Citations on every answer. If the answer isn't in your documents, the bot says so. No hallucinations.
ā Inline figures and tables. When your doc has Figure 3 or Table 2, the bot displays the actual image, not just a description.
ā 12+ languages with full right-to-left support for Arabic, Urdu, and Hebrew. Auto-detects the visitor's browser language.
ā Voice playback on every answer. Bot users can listen instead of read.
ā Image attachments. Bot users can attach a photo, screenshot, or diagram alongside their question.
Tech stack
- Fastify on Node 20 (TypeScript end-to-end)
- Next.js 15 dashboard
- Neon Postgres with pgvector for retrieval
- Voyage AI embeddings, Claude Sonnet for answers
- Cloudflare R2 for document storage
- Stripe billing
- Plain HTML widget (no framework) so embed weight stays minimal
Pricing:
14-day free trial, no credit card. Plans start at $29/month. Cancel anytime.
What's next on the roadmap:
- AI vision descriptions so the bot can find figures even when source docs don't use "Figure N:" captions
- API access for custom integrations
- Multi-team workspaces for larger orgs
Would love your feedback. What would make this useful for YOUR documents? Drop a comment or DM me. I'm watching the thread all day.
Thanks for checking it out š
https://ou1c7z.short.gy/AmH9HO