FigureLabs Review 2026: Credits, Vector Export, and What Users Actually Say

FigureLabs turns a prompt, a sketch, or a reference image into a scientific figure and exports it as SVG or editable PPTX. Checked on its own pricing page on 25 August 2026, its published credit rates charge 50 credits per generation, 60 per text edit, and 150 credits for a single SVG or PPTX export — meaning the export, not the generation, is the most expensive action in the workflow. Plans run from a free tier capped at 1K resolution with no vector export, through Starter at $10/month with 1,000 monthly credits, to Pro at $99/month with 20,000. This review works through the plan structure, the arithmetic for one finished figure, what the vector export does and does not guarantee, and the awkward fact that nearly every "FigureLabs review" currently ranking on Google was written by a competing vendor — including, in the interest of disclosure, this one.

What FigureLabs actually is in 2026

FigureLabs launched in August 2025 and positions itself as an AI agent for scientific illustration rather than an icon library. The home screen offers Illustration, Flowchart, and Data Chart modes; inside a project you get a chat history for revising a figure, reference-image and sketch inputs, palette and visual-consistency controls, upscaling, region redraw, text editing, and a built-in vector canvas. Its own FAQ frames the product as a "BioRender companion": generate a base figure in seconds, export SVG or PPTX, finish the layout in BioRender, PowerPoint, or Illustrator.

Two things follow from that framing, and both are fair. It is genuinely a generator, not an assembly tool — describe a mechanism in a sentence and you get a plausible first composition without hunting through an icon library, which beats drag-and-drop for anything unusual enough that no icon exists. And it is explicitly not trying to be where your figure gets finished; the vector canvas exists, but the FAQ points outward to Illustrator or PowerPoint for "deep editing." Whether that matters depends on how many revision rounds your figures survive.

FigureLabs also runs a multi-model strategy: as of August 2026 the pricing page lists Nano Banana, Nano Banana 2 and Pro, GPT Image 1.5 and 2, SeeDream 4.5 and 5.0 Lite, and Flux.2 Max, gated by tier — two models on Free, five on Starter, eight on Plus and above. Generation costs the same credits whichever model you pick.

The plans, verified as of August 2026

Read from figurelabs.ai/pricing on 25 August 2026. A "Back-to-School Sale" was running at the time — up to 41% off quarterly and 55% off yearly — so the annual column below is a promotional price, not a steady-state one. Prices may change; we re-check monthly.

<table style="min-width: 150px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1">Plan</th><th colspan="1" rowspan="1">Monthly</th><th colspan="1" rowspan="1">Annual (during sale)</th><th colspan="1" rowspan="1">Credits</th><th colspan="1" rowspan="1">Max export</th><th colspan="1" rowspan="1">Vector SVG + PPTX</th></tr><tr><td colspan="1" rowspan="1">Free</td><td colspan="1" rowspan="1">$0</td><td colspan="1" rowspan="1">$0</td><td colspan="1" rowspan="1">150 one-time + 50/day refresh</td><td colspan="1" rowspan="1">1K</td><td colspan="1" rowspan="1">No</td></tr><tr><td colspan="1" rowspan="1">Starter</td><td colspan="1" rowspan="1">$10/mo</td><td colspan="1" rowspan="1">$9/mo ($108 first year, then $120/yr)</td><td colspan="1" rowspan="1">1,000/mo</td><td colspan="1" rowspan="1">4K</td><td colspan="1" rowspan="1">Yes</td></tr><tr><td colspan="1" rowspan="1">Plus</td><td colspan="1" rowspan="1">$35/mo</td><td colspan="1" rowspan="1">$16/mo ($192 first year, then $420/yr)</td><td colspan="1" rowspan="1">5,000/mo</td><td colspan="1" rowspan="1">8K</td><td colspan="1" rowspan="1">Yes</td></tr><tr><td colspan="1" rowspan="1">Pro</td><td colspan="1" rowspan="1">$99/mo</td><td colspan="1" rowspan="1">$45/mo</td><td colspan="1" rowspan="1">20,000/mo</td><td colspan="1" rowspan="1">8K</td><td colspan="1" rowspan="1">Yes</td></tr><tr><td colspan="1" rowspan="1">Team</td><td colspan="1" rowspan="1">$35/seat</td><td colspan="1" rowspan="1">$20/seat</td><td colspan="1" rowspan="1">5,000/seat/mo</td><td colspan="1" rowspan="1">8K</td><td colspan="1" rowspan="1">Yes</td></tr></tbody></table>

Three details in that table deserve more attention than they usually get.

The annual discount is a first-year discount. The Plus plan reads "$192 first year, then $420/year." If you sign up during a sale and forget about it, your second year costs 2.2× your first. Starter goes $108 then $120. Diarise the renewal date the day you subscribe.

The free tier is not a trial of the actual product. It caps output at 1K, excludes vector and PPTX export entirely, and — per the FAQ — is for non-commercial use only. You can evaluate generation quality on it, which is worth doing, but you cannot evaluate the part of the workflow that matters for submission.

Credits do not roll over on individual plans. They reset at the end of each billing cycle. Team and Business plans roll over up to 20%. Paid users also get 100 refresh credits per day, valid only that day, which rewards steady daily use and does nothing for a person who touches the tool twice a month during revision crunch.

The credit table is where the real price hides

FigureLabs publishes its per-action rates, which is more transparency than most tools offer. From the pricing FAQ, as of August 2026:

<table style="min-width: 50px;"><colgroup><col style="min-width: 25px;"><col style="min-width: 25px;"></colgroup><tbody><tr><th colspan="1" rowspan="1">Action</th><th colspan="1" rowspan="1">Credits</th></tr><tr><td colspan="1" rowspan="1">Generate a 1K image</td><td colspan="1" rowspan="1">50</td></tr><tr><td colspan="1" rowspan="1">Regenerate / region redraw / white background</td><td colspan="1" rowspan="1">50 each</td></tr><tr><td colspan="1" rowspan="1">Text edit (fix a label)</td><td colspan="1" rowspan="1">60</td></tr><tr><td colspan="1" rowspan="1">Upscale to 2K / 4K / 8K</td><td colspan="1" rowspan="1">10 / 20 / 40</td></tr><tr><td colspan="1" rowspan="1">Export as SVG or PPTX vector</td><td colspan="1" rowspan="1">150</td></tr></tbody></table>

Nothing about that is hidden or dishonest. But price a realistic figure with it rather than a single lucky generation, and the shape of the plan changes.

What one finished figure costs

A finished figure is rarely one prompt. Call it four generations to get the composition right, two label corrections after your PI reads it, one 4K upscale for the submission portal, and one vector export so a co-author can edit it:

4 × generate/regenerate = 200
2 × text edit = 120
1 × 4K upscale = 20
1 × SVG or PPTX export = 150
Total ≈ 490 credits

Against the plans, that is roughly two finished figures per month on Starter, ten on Plus, and forty on Pro. The absolute floor — one generation, one text fix, one export — is 260 credits.

The structural problem is the ratio. Exporting costs three times as much as generating. That prices the exact behaviour peer review forces on you: every time someone says "can you send me the editable version," you spend three generations' worth of budget. A tool whose most expensive action is the one you repeat during revision is optimised for the demo, not the deadline.

:::figure{size=large align=center}
BLOGFIGUREMARKER0END
:::

Three usage profiles

The occasional user (2–3 figures a year). The free tier plus daily refresh credits genuinely covers exploration. But you cannot export vector or publish commercially, so you will need at least one paid month at submission time. Budget one month of Starter per paper, not a year.

The mid-career postdoc (one paper in revision, four to six figures). This is where Starter breaks. Four figures with real revision rounds is roughly 2,000 credits against a 1,000-credit allowance. You are either upgrading to Plus or buying top-ups mid-deadline, which is the worst moment to be making pricing decisions.

The lab (multiple manuscripts, shared account). Team is $35/seat monthly with 5,000 credits per seat and up to 20% rollover — the only tier with rollover. The arithmetic works if usage is steady, and works badly if your lab's figure production is bursty, clustered around submission and conference deadlines, as most labs' is.

Vector export: what "SVG" does and does not guarantee

"Exports SVG" is the single most over-trusted line on any figure tool's feature list. An SVG file is a container, not a promise. What matters for a reviewer revision is the structure inside it:

Is the text still text? If labels were rasterised into the generated image and then auto-traced, they arrive as outlined paths. You can move them; you cannot retype them. Renaming "IL-6" to "IL-6R" means redrawing letterforms.
Are elements separate objects? A traced bitmap often produces one enormous group of spline shapes. Selecting "the mitochondrion" becomes selecting forty overlapping paths.
Does it open cleanly? Test in whatever you actually use — Illustrator, Inkscape, or PowerPoint — not in a browser preview, which will render almost anything acceptably.

FigureLabs' documented path to editability is vectorisation: vectorise a generated figure, then tweak it in the built-in canvas or export SVG/PPTX for deep editing elsewhere. That is a legitimate workflow and for many figures it is enough, but it is structurally different from a pipeline that keeps labels as live text objects from the start. Before committing a year of budget, run this test: generate one figure of the kind you actually publish, export the SVG, open it in Inkscape, and try to change one label.

Its FAQ states that 8K raster output corresponds to roughly 1200 DPI and that vector output suits posters and large prints. Both are reasonable claims, but effective DPI depends on final printed dimensions — check against your target journal's size spec, not the number alone.

What users actually say — and why that is hard to answer honestly

Here is the uncomfortable finding. Searching for FigureLabs reviews in August 2026 returns almost nothing written by an unaffiliated researcher. There is no substantial G2, Capterra, or Trustpilot profile. There is no significant Reddit thread in the academic subreddits. What ranks instead are articles published by competing figure tools — SciDraw AI and PaperBanana both run "FigureLabs alternative" pages that double as reviews, and FigPad, whose team wrote this article, publishes a comparison page too. Treat all of us accordingly.

Reading those competitor reviews against the vendor's own pages, three criticisms recur and two of them check out:

Post-generation editing is limited without leaving the tool. This one is corroborated by FigureLabs' own FAQ, which routes deep editing to Illustrator, BioRender, or PowerPoint. It is a design choice, not a defect, but it is real.

Team pricing scales sharply. Also consistent with the published page: $35 per seat per month is not a volume discount over the $35 individual Plus plan; you get rollover and shared credits instead.

The free tier watermarks output. This one we could not verify. FigureLabs' current pricing page does not mention a watermark; it restricts the free tier to 1K resolution, no vector export, and non-commercial use. If watermarking matters to your decision, test it yourself rather than trusting any secondhand claim, ours included.

On the positive side, the coherence of the workspace — prompt history, reference images, revision actions, and canvas in one place — is the thing even competing reviews concede. If your bottleneck is getting from a blank page to a defensible first composition, that is a genuine strength.

How FigPad fits

We build FigPad, so read this section as an argument rather than a finding. FigPad exists for the same moment FigureLabs does — prompt, sketch, or reference image in, scientific figure out — with one deliberate difference in where the cost sits. Vector SVG and PPTX export is a plan feature from the Plus tier upward rather than a per-export charge, so re-exporting after a reviewer comment does not compete with your generation budget. FigPad's own pricing FAQ puts the same generate–revise–export loop described above at roughly 16 credits on its side; that is our arithmetic about our own product, so verify it on a figure you actually need before believing it.

The other difference is what comes out. FigPad exports layered SVG in which labels stay independently editable text objects, plus editable PPTX for the co-authors who will never open a vector editor, and 2K–8K PNG depending on plan. A built-in vectorizer converts an existing bitmap — a screenshot, an old slide, a scanned sketch — into editable paths. Paid plans include publication rights, and the terms state that outputs belong to you and that uploads are not used to train models.

Practical entry points: if you want the direct feature-by-feature layout against your current subscription, the FigureLabs alternative page has it. If you need a figure today, the AI Scientific Figure Generator is the shortest path from a methods paragraph to a first draft, and the Graphical Abstract Generator handles journal graphical-abstract specs. Current tier details are on the pricing page. According to FigPad, 100K+ researchers use it, at institutions including MIT, Stanford, UC Berkeley, Yale, Cornell and Cambridge. One honest limit: the PDF-to-poster feature accepts files up to 80 MB and parses only the first 30 pages, which covers a manuscript but not a thesis.

When you should stay with FigureLabs

Switching mid-project costs more than most people estimate. Several situations argue for staying.

You want model choice in one interface. Eight image models on a single account, all costing the same credits per generation, is a real feature. If your practice is generating the same prompt across Nano Banana Pro, GPT Image 2, and Flux.2 Max and picking the best, no single-model tool replaces that.

You log in most days. The 100 daily refresh credits are worth about two generations a day, every day, on top of your plan allowance. For someone who works on figures continuously, the effective cost per figure is much better than the monthly arithmetic suggests.

You are already on Pro or Team. At 20,000 credits a month, a 150-credit export stops being a constraint worth reorganising your workflow around.

Your history lives there. Months of iterations, reference sets, and team folders have genuine switching cost. Do not move a manuscript mid-revision to save $20.

The same fairness applies to BioRender: if your figures are cell or molecular biology and your reviewers expect its conventions, its icon library remains the best in the field and no AI generator has matched that standardisation.

Key Takeaways

The export is the expensive action. At FigureLabs' published August 2026 rates, an SVG or PPTX export costs 150 credits — three times a generation — which prices exactly the behaviour peer review demands.
One finished figure is roughly 490 credits, so the $10/month Starter plan's 1,000 credits buy about two per month. Compare tools on cost per finished figure, not price per month.
The annual discount is a first-year discount. Plus reads "$192 first year, then $420/year." Set a calendar reminder before the renewal, not after.
The free tier cannot test the thing that matters — it is capped at 1K, excludes vector export, and is non-commercial only.
"Exports SVG" is not a promise of editability. Generate one real figure, export it, open it in Inkscape, and try to rename a label. That test outranks every review, including this one.

👉 Start for free — no credit card