EasyBI developer documentation
EasyBI (easybi.ai) is a web application for analyzing spreadsheets and documents. This page summarizes the product for developers, AI agents, and anyone evaluating it without creating an account.
What EasyBI does
You upload a spreadsheet or document, ask questions about it in plain language, and receive charts, tables and written answers. Every account starts with a 14-day free trial; no credit card required.
- Uploads: CSV, Excel (XLSX), PDF and Word (DOCX).
- Questions are asked in chat, in everyday language.
- Answers are generated by AI-written analysis code executed against your real data.
- Results can be rendered as interactive charts and tables inside the conversation.
How an analysis works
Understanding the pipeline explains what the product stores and what leaves your systems:
- 1. Upload — a file is stored in blob storage and normalized into a canonical JSON dataset (sheets with columns, rows and inferred column types).
- 2. Catalog — the model sees only file names, column names, row counts and a short generated summary per file — never full raw rows.
- 3. Chat — you ask a question; the model may request an analysis run.
- 4. Run — EasyBI writes Node.js code, runs it in an isolated sandbox with no network access, with only your selected datasets mounted.
- 5. Result — the run produces a result table (up to 5000 rows); the model sees a preview of at most 20 rows, and the UI renders the full result as a chart or table.
Plans and limits
Prices in US dollars per month, billed through Stripe. The trial needs no card.
| Plan | Price | Questions/mo | Document pages | File uploads | Storage |
|---|---|---|---|---|---|
| Trial | Free | 50 | 200 | 15 | 5 GB |
| Starter | $5/mo | 100 | 400 | 30 | 2 GB |
| Standard | $10/mo | 200 | 750 | 60 | 5 GB |
| Pro | $20/mo | 750 | 1500 | 120 | 15 GB |
Machine-readable surfaces
Everything an agent needs to represent EasyBI is public and fetchable without JavaScript or authentication:
- /llms.txt — curated index of LLM-readable pages (llmstxt.org format).
- /sitemap.xml — search-engine page list; /robots.txt — crawl rules.
- Markdown negotiation — request any page on this site with Accept: text/markdown to receive its Markdown representation; each page also has a .md sibling URL.
- /api/health — unauthenticated liveness endpoint returning {"status":"ok"} (or 503 {"status":"degraded"} when the database is unreachable).
Product API
The application is driven by a session-cookie-authenticated JSON API under /api (files, runs, chats, projects, billing). It authenticates the browser session of a signed-in user; it is not yet a public partner API and offers no API keys. Agents should use the machine-readable surfaces above instead of the product API.
Privacy position
Model requests are sent only to providers operating under zero data retention and no-prompt-training terms. Spreadsheets are parsed on our own servers; PDF/Word extraction sends the document to Google Gemini for table extraction. We never sell data or train models on it.
Contact
Privacy questions and account deletion requests: foad@fmktech.com.br. Full policies: /privacy and /terms.