FigCanvas
FigCanvas is an AI figure generator for researchers. It turns a plain-language prompt, a methods section or a spreadsheet into publication-ready illustrations, charts and flowcharts, with daily free credits and no card required.
What is FigCanvas?
FigCanvas is a browser-based figure generator that positions itself as AI for Scientists, and it brings together three production lines researchers usually have to stitch from separate tools: scientific illustration, data visualisation and flowcharting.
On the illustration side, it draws signalling pathways, mechanisms, cell biology schematics, experimental overviews and graphical abstracts from a written description. On the data side, you upload a CSV, TSV, TXT or Excel file of up to 10 MB; the tool reads it, recommends a suitable chart type and generates the plot. The families covered are the ones a life-science paper actually needs: differential expression volcano plots, heatmaps and expression matrices, statistical comparisons with bar, box and violin plots, correlation and dimensionality reduction with scatter plots and PCA, survival and clinical outcomes, and forest plots for meta-analysis, alongside Sankey diagrams, ridge plots and growth curves. On the diagram side, you paste a methods section or a protocol and the AI extracts the steps into an editable flowchart; worked examples on the site cover PRISMA, CONSORT, clinical trial and factorial RCT randomisation, a CNV-calling bioinformatics pipeline, a microscopy image processing pipeline and an ML feature engineering pipeline. The menu names these entry points individually: Scientific Illustration, Graphical Abstract Maker, Data Visualization, Flowchart, PRISMA Diagram Generator, Volcano Plot, Forest Plot and a Scientific Color Palette Generator.
Two design decisions matter as much as the generation itself. The canvas is infinite, so charts, illustrations and diagrams sit side by side with smart alignment and grouping guides, and a single plot grows into a six-panel manuscript figure exported as one composition. And the output is not flattened: right-click a figure, choose Vectorize, then Edit SVG, and labels, colours and individual elements can be corrected in the browser without third-party software.
The editorial stance is explicit. Palettes, typography and conventions such as error bars, axis labels and significance markers are calibrated on Nature, Science and Cell, the publisher stating that We studied thousands of figures from top-tier journals to learn what publication-ready actually means. The models are described as fine-tuned on scientific literature and datasets rather than general-purpose image generators. Marketing claims 20,000+ researchers and a 100,000+ template library. Access is browser-only with a Google or email account; there is no mobile app and no API.
What it does
- Generate a scientific illustration from a plain-language description
- Turn a CSV or Excel dataset into a publication-ready chart
- Convert a methods section into an editable flowchart
- Produce a PRISMA 2020 diagram for a systematic review
- Assemble a multi-panel figure on an infinite canvas
- Vectorise a generated figure and edit it element by element
- Export to 300 DPI PNG, editable SVG or PDF
When to use FigCanvas / When not to
A quick filter to help you decide if FigCanvas is the right fit.
When to use FigCanvas
- Researchers who have to deliver a manuscript-quality figure without any design training
- PhD students and postdocs preparing papers, conference posters and thesis defences
- Systematic reviewers and clinical teams who need PRISMA 2020, CONSORT, enrolment flow and randomisation diagrams
- Omics and bioinformatics groups plotting volcano plots, heatmaps, PCA, forest plots and survival curves
- Labs with no access to a scientific illustrator, and anyone unwilling to write code just to draw a chart
When not to use FigCanvas
- Developers who need an API: the FAQ states plainly that FigCanvas is designed for researchers, not developers
- Anyone working with identifiable personal or patient data, since no GDPR claim, no DPA and no security certification appear anywhere on the site
- Users who need vector output for free: the free tier exports watermarked PNG only, capped at three projects and five cards per project
- Mobile-first users, as there is no iOS or Android app and no browser extension, only a web app
- People looking for a biomedical icon library to assemble by hand, or for a general-purpose layout and presentation suite
How to use FigCanvas
A typical end-to-end flow, from setup to results.
- Create an account with Google OAuth or an email address; daily free credits start immediately and no card is required
- Provide an input: a plain-language prompt, a methods section or protocol, or a CSV, TSV, TXT or Excel file of up to 10 MB - sample datasets such as clinical_trial.csv, rnaseq_degs.csv, metabolomics.csv and group_comparison.csv are available
- Let the AI select the figure type that fits the input and produce a first draft, announced in under two minutes
- Refine the result by rephrasing the prompt and correcting labels and layout; data visualisations and flowcharts are editable directly on the canvas
- For an illustration, right-click and choose Vectorize, then Edit SVG to retouch text, colours and individual elements in the browser
- Compose a multi-panel figure on the infinite canvas by dragging, aligning and grouping with the smart guides
- Export to PNG (300 DPI on Pro), editable SVG or PDF; the free tier produces watermarked PNG only
- Keep an eye on the credit balance: every generation costs credits at a rate that depends on the job - quick plot, high-quality plot, diagram, high-quality AI image or vectorisation
Pros & Cons
Pros
- Three figure types - illustration, data visualisation and flowchart - in a single tool, where researchers usually have to combine several
- No coding and no design skills required: the starting point is natural language, an example or a data file
- Output stays editable, with SVG, text layers and element-by-element retouching in the browser rather than a flattened image
- Full ownership of the generated figures, royalty-free and publishable in journals with no additional licence
- A written, unambiguous position on training: uploaded datasets and private prompts are not used to train the publisher's public models, and uploaded datasets are deleted within 24 hours unless saved to a project
- A permanent free plan with full publication rights and no card required, and a low paid entry point compared with the sector
- Academic use cases named outright - PRISMA, CONSORT, RCT randomisation, bioinformatics pipelines - with journal conventions applied by default
Cons
- GDPR and CCPA are mentioned nowhere on the site: no customer DPA, no Article 27 representative, no named DPO, and no formal subprocessor list beyond the four vendors cited in the privacy policy
- No corporate identity is published: no full legal name with a legal suffix, no postal address, no identifiable founder or beneficial owner
- No security certification is displayed, neither SOC 2, nor ISO 27001, nor HIPAA
- A very young product: the domain was registered on 30 March 2026 and first archived on 16 June 2026
- No API, no mobile app and no browser extension, and the only support channel is a generic email address
- A tightly capped free tier: watermarked PNG only, no SVG or PDF, three projects and five cards per project
- Credits never roll over from one cycle to the next and pack credits expire after 30 days, while the terms leave scientific accuracy entirely to the user
Pricing & Plans
A permanent free plan is available at no cost, and the cheapest paid entry point is the Pro plan at USD 19.90 per month. An annual commitment brings it to USD 149.90 billed as a single payment, which the publisher presents as the equivalent of USD 12.50 per month. One-off credit packs and a quote-based Custom plan complete the range. The free tier grants 10 credits per day, raised to 18 per day during the first three days, while Pro grants 1,000 credits per month - the publisher citing 60 to 200 figures per month as typical Pro usage. Credits do not roll over, subscriptions renew automatically at the end of each cycle unless cancelled from the billing portal, and refunds are handled case by case under section 8 of the terms.
- 18 credits per day for the first three days then 10 per day
- access to the 100
- 000+ template library
- publication-ready illustrations
- editable flowcharts and diagrams
- no-code data visualisation
- the interactive canvas editor
- up to 3 projects and up to 5 cards per project
- 1
- 000 credits per month
- unlimited projects
- unlimited cards per project
- full PNG
- SVG and PDF export without watermark
- access to high-quality generations
- priority support
- the same features as Pro Monthly
- the same features unlocked
- with watermark-free exports and code downloads charged at 5 credits each
- pack credits expire after 30 days
- tailored credit allocation
- team management
- volume discounts
- invoice billing
- dedicated support
- custom integrations
- site licences and SSO
Data, GDPR & hosting
A consolidated view of how FigCanvas handles your data.
GDPR overview
There is no GDPR implementation to report. The words GDPR and CCPA appear nowhere on the site, neither in the privacy policy nor in the terms. No legal basis is cited, no data protection officer is named and no Article 27 representative is designated; no DPA is offered to customers, the only mention of data protection agreements covering the publisher's own vendors. What does exist is a rights section written in GDPR-adjacent vocabulary: You can access, correct, or delete your personal data at any time. You can export your data or request that we restrict how we process it. The single channel is contact@figcanvas.com. On transfers, data may be processed in the United States and the European Union under unspecified appropriate safeguards. The stated age floor is 13, the American COPPA threshold rather than a European one.
Who owns the data?
Section 5 of the terms is unambiguous: You own your uploaded data, text prompts, and the projects you create. FigCanvas takes only a limited licence to store, process and display that content, strictly to run the platform, and the licence ends when the content or the account is deleted. Section 6 goes further: You retain full ownership of AI-generated figures created through FigCanvas, royalty-free and usable for any lawful purpose, journal publication included. The exception sits in section 7: reference figures taken from open-access literature are displayed for guidance only, are excluded from exports and may not be reused as your own work. Projects are private by default.
Reuse rights
You can reuse what you create without asking permission: generated figures are royalty-free and can be published in Nature, Science, Cell, ACS journals or any academic outlet with no additional licence. On the publisher's side, the privacy policy effective 6 April 2026 lists what is collected: account details (email, display name, profile picture URL from Google OAuth or email sign-up), usage data (pages visited, features used, device and browser type, IP address, approximate location) and the content you create - projects, canvas files, uploaded datasets, prompts and generated figures. That material serves three declared purposes: running the service, improving the product and the quality of AI output, and sending service and security messages. Only prompts and data summaries, not full datasets, reach the named AI providers - Third-party AI providers (Anthropic, Google, OpenAI) for figure generation, receiving only prompts and data summaries - while E2B runs code in isolated sandboxes. The policy commits that We do not sell, rent, or trade your personal information. and the product FAQ adds that uploaded datasets and private prompts are not used to train the publisher's public models. No training opt-out is documented, the publisher treating the question as moot.
Data retention & training
Hosting summary
The privacy policy says only that data may be processed in the United States and the European Union. No region, data centre or hosting provider is named, no residency choice is offered to the user, and the appropriate safeguards invoked for international transfers are never spelled out. On protection the policy is more concrete: All data is transmitted over HTTPS and stored encrypted., backed by regular security audits, access controls and rate limiting. Code execution for data visualisation is confined to isolated sandboxes operated by E2B, while prompts and data summaries are passed to Anthropic, Google and OpenAI for generation. One technical observation completes the picture without being a statement from the publisher: the domain resolves to 188.114.97.0, an anycast address on AS13335 (Cloudflare), which describes the edge in front of the site rather than where user files are stored. In short, a US-leaning setup with a European processing possibility, described in general terms and unsupported by any published hosting commitment.
Things to keep in mind
Risks and trade-offs to weigh before adopting FigCanvas.
- No GDPR compliance is claimed anywhere on the site, no customer DPA is offered and neither an Article 27 representative nor a DPO is named - a real obstacle for a European lab handling personal data
- The publisher stays anonymous behind a bare footer: no full legal name, no postal address, no identifiable founder or beneficial owner, and no security certification, for a service online for less than a year
- The terms put the burden of correctness on you: You are responsible for verifying the accuracy of AI-generated content before publishing - a figure that looks like Nature is not the same as a figure that is right
- The AI can return different results for the same input, so a figure is never exactly reproducible; every label, axis and value has to be checked before submission
- The terms recommend noting that figures were created with AI assistance when publishing, in line with emerging journal policies, and reference figures from the literature must never be exported or presented as your own work
- Leaning on the generator for every figure can quietly erode the habit of reasoning a figure through - which comparison, which scale, which encoding - and that skill is hard to get back
- Commercial fine print to keep in mind: credits do not roll over, pack credits expire after 30 days, subscriptions renew automatically, and the stated age floor is only 13
Setup & Integrations
Technical difficulty
Very low. There is nothing to install, since FigCanvas runs entirely in the browser. You need an account, created through Google OAuth or an email address, and nothing more: no connector to configure, no environment to prepare and, as the site repeats on every page, no coding skills. The only technical constraint is the data file itself, which must be CSV, TSV, TXT or Excel and no larger than 10 MB. The learning curve starts from natural language, a worked example or your own dataset, and the publisher announces a first draft in under two minutes.
Deployment
Integrations
Supported languages
Behind FigCanvas
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement FigCanvas.
Frequently asked questions
What is FigCanvas?
Is FigCanvas free to use?
What types of figures can I create?
Can I publish FigCanvas figures in a journal?
Which export formats are available?
Do I need to know how to code?
Can I edit a figure after it has been generated?
Are my uploaded data used to train the AI?
How long are my uploaded files kept?
Do unused credits roll over?
Should you pick FigCanvas?
FigCanvas is a young but well-aimed tool. It covers three figure needs - illustration, data visualisation and flowcharting - in one place, without asking researchers to learn a design suite or write plotting code. Its two strongest cards are output that stays editable, down to SVG elements and text layers, and publication rights that are complete from the free tier upwards: the figures are yours, royalty-free, usable in Nature, Science, Cell or any academic outlet.
The structural weakness is legal opacity. The site publishes no full company name, no postal address, no founder and no beneficial owner; GDPR is not mentioned once, no DPA is offered and no security certification is displayed. Nothing here points to bad faith - the privacy policy is clear about what is collected and states that uploaded datasets do not train public models - but a European institution running a procurement check will find very little to hold on to. The product is also very recent, the domain registered in March 2026 and first archived in June 2026, which is under a year of visible existence.
The practical reading is straightforward. For non-identifying research data - expression matrices, trial flow counts, protocol steps, pathway schematics - FigCanvas is a reasonable and inexpensive choice at USD 19.90 per month, and the permanent free plan is generous enough to judge output quality before paying anything. For identifiable personal or patient data, it should be avoided until the publisher documents a GDPR position and offers a data processing agreement. And whatever the tier, the terms are right to insist on it: the scientific accuracy of a generated figure remains the author's responsibility, and journals increasingly expect AI assistance to be declared.
- Choosing a selection results in a full page refresh.
- Opens in a new window.