Charts you can actually publish.
Describe the chart you want. ChartForce routes it to a primary source — a government agency, a central bank, a statistical office — fetches the real numbers, and bakes the citation into the file you export. The model plans the query. It never types a number.
Most "AI charts" are
quietly making things up.
Ask ChatGPT for "US GDP from 2010 to 2024" and it will hand you a confident table of numbers. Most of them will be wrong. None of them will be cited.
ChartForce is built on one rule: the AI plans, the API delivers. Models route the query — they never invent a number.
Primary sources only
Government agencies, central banks, intergovernmental bodies. No scraped Wikipedia. No aggregator-of-aggregators.
The model never types a number
The AI parses your prompt and routes it to the right dataset. Numbers enter the system only after a real fetch from a real API, and only after a validation gate.
Every chart cites itself
Source name, dataset ID, retrieval timestamp — baked into the export. APA, MLA, plain-text. Nothing you publish needs to be back-filled.
Bad data fails loud
Missing values flagged. Outliers surfaced. If the data isn't clean, the job fails with an explanation. Better than a confidently-wrong chart.
From a sentence to a sourced chart, in five stages.
Most jobs finish in under a minute. You can watch every stage as it runs.
Describe
Plain-English prompt. No SQL, no spreadsheets.
Route
AI plans which sources to query — FRED, WB, Census, BLS, IMF.
Fetch
Real API calls. Real files. Numbers enter the system here.
Clean
Normalize units, fill gaps, flag outliers. Validation gate.
Render
Chart type chosen by data shape — including animated races. Citations baked in, PNG, SVG or MP4 out.
Charts ChartForce actually makes.
Three exports, three prompts, three sources. Each card shows the prompt that produced the chart and the citation that ships with it.
A static chart gets scrolled past.
The same sourcing, the same citation, the same validation gate — playing. Every number in a race was fetched from a primary API, which is the part that makes it publishable rather than just watchable.
- Bar races and line races, from the same prompt
Ask for a ranking over time and you get a race, not a static snapshot of the last year. The chart type is chosen from the shape of the data, the same as everything else here.
- Exported as MP4, up to 60fps
Encoded in your browser, so nothing queues and nothing uploads. Free exports at 1080p and 30fps; Pro and Max at 60.
- Sized for where it is going
Story and Reel at 1080×1920, square at 1080×1080, landscape for X, 16:9 for a deck. The preset follows what you said the chart was for.
- Faces on the bars
Upload a logo or a portrait once and it attaches to that entity in every race that mentions it — matched on the name, never guessed at.
Why not just ask ChatGPT?
Because ChatGPT can give you a confident chart of made-up numbers. ChartForce exists to fix that.
ChartForcePeople whose credibility is on the line.
Real data, fair pricing.
Every plan runs the full pipeline. You're paying for output volume — not for honest data.
Try the engine. Ship a few charts.
- Monthly credit allowance
- PNG + animated MP4 · watermarked
- 1080p · 30fps video
- Citations included
Built for creators and analysts who publish weekly.
- 10× the monthly credits
- 60fps MP4 · no watermark
- PNG + SVG + MP4 export
- Screenshot → data where enabled
Volume output for studios and content teams.
- Max monthly credits
- Everything in Pro
- 8K export bounds
- Priority support
Buying for a team or a company? Talk to us about volume, seats and invoicing.