Image and video decision
Recraft
A capable developer can reproduce a useful subset (prompt→generate→store→basic editor→billing) by integrating external models and open-source editor components, but you cannot reproduce Recraft’s claimed proprietary model quality or built-in vector-generation fidelity without significant investment.
Visit website↗No open-source build does this yet
Nothing published replaces this one, so a replacement starts from an empty file. Here is what it would have to cover.
What a replacement has to do
- Accept text prompt + optional style images → generate raster or editable SVG vector → store result, present in web editor, allow edits/exports, bill for usage
What it still won’t have
- Proprietary Recraft model quality and any unique art-direction baked into their models
- Native editable-vector generation parity (quality and fidelity)
- Enterprise SLA, managed hosting, and integrated product polish
- Any private features tied to Recraft’s internal models or datasets
What remains hard
- Proprietary models
Recraft has released four generations of its own model:
- Proprietary models
Meet Recraft V4.1: Our most advanced model. Beautiful by nature.
First-year cost
No published price
Recraft does not publish a price we could read, so there is nothing to compare against. What building costs is below.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal Recraft-like service using Next.js + React frontend, Node.js backend (Express), Postgres for metadata, S3-compatible storage for assets, and Redis+BullMQ for async jobs. Scope: (1) web UI to submit text prompts and upload reference images, (2) backend job worker that calls an external image model API (e.g., Stable Diffusion/Replicate) and a vectorization step to produce SVG, (3) store outputs and metadata, (4) simple web editor to crop/replace background and download PNG/SVG, (5) per-user credit counter and Stripe subscription for 1 seat. Out of scope: training new generative models, advanced art-direction models, enterprise SSO. Include error handling for failed model jobs, retries, and automated tests for API endpoints and worker jobs.
How we checked
How the score was reached
- Partly verdict base52
- 3 cited sources+3
- Hard moats found in the evidence-3
- Evidence score52
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time — so the same evidence always produces the same number.
How scoring works →Cited sources · 3
Every page the run actually retrieved.
- official productRecraft home
- official pricingRecraft pricing
- official docsRecraft docs — models
Integrity checks
What held up, and what did not.
