Pixel Dojo
Pixel Dojo is a single web workspace bundling 70+ third-party AI models for images, video and audio. Creators, marketers and agencies generate, edit, animate and upscale under one subscription with full commercial rights, from $10 per month.
What is Pixel Dojo?
Pixel Dojo is a web platform for AI visual creation that aggregates third-party models rather than running its own. The pitch is arithmetic: instead of paying one subscription per model, you pay one subscription for all of them. The homepage prices the alternative at $164 a month (Midjourney $30, Runway $35, Magnific $39, Luma $30, Leonardo $30) against $25 for the Pro plan.
The homepage advertises 70+ models; the API platform lists 153 (86 image, 63 video, 3 audio, 1 3D). Names include FLUX, Seedream, Seedance, Kling, VEO 3.1, WAN, Qwen, Nano Banana, GPT Image 2, Ideogram 4, Grok Imagine, Recraft, Hunyuan, MiniMax, Pixverse, Vidu, Magnific, ElevenLabs Music, Heygen and OmniHuman.
Above the raw models sit three Studios, included in every paid plan. Film Studio writes a script from a sentence, storyboards each shot, renders clips with lip-synced dialogue and exports an edited film. Marketing Studio turns a product URL into a voiced video ad in one of 24 styles. Character Studio stores references for characters, casting, locations, products, outfits, props and styles, and produces shoot boards. Around them: a multi-tool Canvas, a community gallery, contests, a preset library, a prompt builder and per-model prompting guides. LoRA training is included from the Pro plan up.
Everything is metered in credits. Each model shows its credit cost before you run it; plans grant 160, 400 or 800 credits per month. The same subscription therefore yields a very variable number of generations, since costs range from 0.1 to 45 credits.
There is a developer route too: a public REST API with API keys, prepaid credits, OpenAPI 3.1, agent-facing docs and a usage dashboard, plus a hosted MCP server with 18 documented skills, usable from Claude Code, Claude Desktop, Cursor, Codex, Windsurf, Zed and Cowork. Zapier and ComfyUI integrations are declared.
The publisher claims 4,000+ active creators, a 4.9/5 rating and 4M+ creations, and documents four safety layers: prompt analysis, an image classifier before submission, a Cloudflare and NCMEC CSAM scan at storage, and provider-side filters. The operator is BLOVE INC, a Colorado S-Corporation in Colorado Springs; Colorado law and Denver County courts govern. The service is active, with a changelog running since April 2025.
What it does
- Generate images from a plain-language prompt across 86 image models
- Generate and animate video with 63 video models, several of them with synchronised audio
- Edit an existing image in place: inpainting, outpainting, relighting, camera-angle changes, upscaling to 4K and up to 16x
- Train custom LoRAs and keep a character consistent from one scene to the next
- Produce a multi-shot short film from a single sentence: script, storyboard, lip-synced dialogue, edited export
- Turn a store URL into a voiced video ad chosen from 24 styles
- Generate voice, music and full audio scenes, and drive all of it from a REST API or an AI agent
When to use Pixel Dojo / When not to
A quick filter to help you decide if Pixel Dojo is the right fit.
When to use Pixel Dojo
- Creators and freelancers juggling several paid model subscriptions who want image, video and audio generation under a single plan
- Marketing and e-commerce teams that need product ads fast: Marketing Studio turns a store URL into a voiced video in one of 24 styles
- Art directors and illustrators who need a recurring character to stay consistent from one scene to the next, through Character Studio
- Founders and small agencies producing a high volume of visuals without an in-house design or motion team
- Developers and agent builders who want a public REST API plus a hosted MCP server callable from Claude Code, Cursor or Windsurf
When not to use Pixel Dojo
- Buyers who need contractual data guarantees: no GDPR commitment, no DPA, no Article 27 EU representative and no SOC 2 or ISO 27001 certification are offered
- Anyone producing adult material: pornography, sexually explicit content and graphic nudity are banned by the terms, artistic nudity included
- Teams working in safety-critical fields, since the terms exclude medical, life-support, autonomous vehicle, aviation and nuclear contexts
- Resellers and white-label operators: reselling, sublicensing and multi-tenant access are prohibited without a written agreement
- Mobile-first users and organisations that must pick their hosting region: there is no iOS or Android app and no regional hosting choice
How to use Pixel Dojo
A typical end-to-end flow, from setup to results.
- Create an account: registration is required before you can use the service
- Pick a plan (Lite, Pro or Pro Max) and pay through the partner or reseller acting as merchant of record
- Open a tool or a Studio, write a prompt in plain language, keep the default settings and refine only when you want more control
- Check the credit cost, which is displayed before the generation runs
- Click Save or Download to keep an image: session results stay in your browser's local storage for one hour only
- In Film Studio, type one sentence and let it write the script, storyboard each shot, render lip-synced clips and export an edited film
- In Marketing Studio, paste a store URL, choose one of the 24 ad styles and collect a video with voice-over
- In Character Studio, register a character once, pin the best result and reuse it across the following scenes
- For agents, run the terminal command claude mcp add --transport http pixeldojo https://pixeldojo.ai/mcp then approve in the browser at first use, with no API key to copy; on claude.ai go to Settings, Connectors, Add custom connector and paste the MCP URL
- For code, create an API key, buy prepaid credits, read the documentation and call /api/v1/models/{model}/run; cancel whenever you want, the plan staying active until the end of the period already paid for
Pros & Cons
Pros
- One subscription covers every model and every Studio, new releases included
- Rare catalogue breadth: image, video, audio, 3D, upscaling and editing all in the same place
- Low entry point at $10 per month with no commitment, cancellation at any time and a full refund on a first purchase within 7 days
- The credit cost is shown before a model runs, and credits are only charged when the generation succeeds
- Full commercial rights on outputs, with an explicit assignment written into the terms
- A public REST API with OpenAPI 3.1 in every plan, plus a hosted MCP server that needs no API key to copy
- Operational transparency: a subprocessor page naming each provider with purpose, data types and location, a safety system described layer by layer with reports handled within 5 business days, session images not persisted by default, and abundant documentation including per-model prompting guides and generation-time comparisons measured on the same prompt
Cons
- No GDPR mention anywhere on the site, no DPA, no Article 27 representative and no certification claimed such as SOC 2 or ISO 27001
- Training on customer data is only ruled out without your explicit consent: the consent mechanism is undocumented and no opt-out is described
- Whether a free tier exists is unclear, since the homepage JSON-LD announces a free tier with included credits that the pricing page does not show
- No mobile app, no about page and no contact page: one email address handles support, legal, privacy and abuse, and no team member or executive is named
- The credit system takes learning, the same subscription yielding wildly different volumes since a generation costs between 0.1 and 45 credits, and third-party model availability or credit pricing can move
- NSFW is prohibited, artistic nudity included, and the customer gets no choice of hosting region
- Beyond the first purchase nothing is refunded and consumed credits never are, while one inference provider, Replicate, processes prompts, images and video frames in the United States
Pricing & Plans
Pixel Dojo is sold as a paid monthly subscription. The pricing page displays no free plan, and the lowest published entry point is USD 10.00 per month for Lite, followed by USD 25.00 per month for Pro and USD 50.00 per month for Pro Max, granting 160, 400 and 800 credits respectively. The Pro plan may also be committed for six months at USD 125.00 or for one year at USD 250.00. Every plan includes the 70+ models, the three Studios, Canvas and the public API; LoRA training is listed on Pro and Pro Max, and Pro Max adds a priority queue on busy days. API credits are bought separately on a prepaid basis. The credit cost per model ranges from 0.1 credit (Flux 2 Klein 4B, P-Image) to 45 credits (OmniHuman), with video often billed per second. Subscriptions may be cancelled at any time, access is immediate, and a first purchase is fully refundable within 7 days, possibly prorated against the compute consumed; credits carry no monetary value and are not refundable once consumed. The publisher's own comparison sets USD 164.00 per month for Midjourney, Runway, Magnific, Luma and Leonardo combined against USD 25.00, a claimed saving of USD 139.00. No enterprise plan and no quote-based pricing are displayed.
- 160 credits per month
- the 70+ models
- Film
- Marketing and Character Studios plus Canvas
- the public API and the community gallery
- 400 credits per month
- all models including every new release
- all Studios plus Canvas and LoRA training
- with commitment options at USD 125 for six months or USD 250 for one year
- 800 credits per month
- all Studios
- Canvas and all models
- plus a priority queue on busy days
- Prepaid API credits
- purchased separately from any subscription on the Buy Credits page
Data, GDPR & hosting
A consolidated view of how Pixel Dojo handles your data.
GDPR overview
Pixel Dojo makes no GDPR claim. The acronym appears nowhere on the site, in text or HTML: no compliance statement, no published or offered DPA, no Article 27 EU representative, no named DPO. The privacy policy, effective 1 November 2025, nevertheless borrows the regulation's vocabulary. It names a controller (BLOVE INC), states purposes and legal bases, lists recipients, and grants data-subject rights: access, rectification, erasure, objection, restriction, portability, withdrawal of consent, and the right not to be subject to a solely automated decision. It acknowledges international transfers, notably to the United States, under Standard Contractual Clauses or adequacy decisions. Rights are exercised through one channel only, support@pixeldojo.ai. The service is not directed to children under 13, or 16 in certain jurisdictions. Read this as GDPR-shaped drafting rather than a stated commitment.
Who owns the data?
The terms split ownership cleanly. Input stays yours: you keep your rights and grant the publisher only a limited licence to process, store, transmit and display your Assets, solely to provide the Service. Output is assigned to you, the publisher transferring whatever rights it holds in content generated for your account, subject to third-party rights and to the terms; the homepage and FAQ confirm full commercial rights over the images and videos produced. The models and software themselves remain the publisher's property, with reverse engineering and model extraction prohibited. BLOVE INC, a Colorado S-Corporation, acts as the data controller behind the service.
Reuse rights
Outputs can be reused commercially without asking further permission, subject to each model's own restrictions and to applicable law. On the publisher's side, the privacy policy effective 1 November 2025 lists the categories collected: identification and account data, economic and transactional data, uploaded or generated text and media together with their metadata, device and activity data such as IP address, browser and pages viewed, support exchanges, and cookies. No full card number is stored, only billing metadata (invoice ID, last four digits, expiry, country, tax status). Stated purposes are delivering and managing the service, product improvement and fraud prevention, commercial information, and responding to authorities, on the legal bases of contractual performance, legitimate interest or consent depending on the case. The policy states that prompts and outputs are not used to train the publisher's proprietary models without explicit consent, while third-party providers may process content to deliver features. Images generated during a session live only in the browser's local storage for up to one hour and are cleared unless you click Save or Download. Risk scoring, IP reputation and behavioural modelling serve fraud prevention, automated and manual moderation inspects prompts and images (Llama Guard 3 and 4, Falcons AI NSFW, a banned-phrase list with contextual GPT-5 review), and outputs may carry watermarks, metadata or safety labels.
Data retention & training
Hosting summary
Pixel Dojo publishes a subprocessor page naming nine providers with their purpose, the data they handle and their location. DigitalOcean runs hosting and infrastructure (IP addresses, operational logs) in the United States and the EU. Replicate performs inference and therefore processes prompts, images, video frames and metadata in the United States. Cloudflare provides CDN, security, DDoS and anti-bot services from a global edge, handling IP addresses and request metadata. MongoDB Atlas is the main database (account, usage, logs) in the United States and the EU, while PlanetScale holds account and usage metadata in the United States. Brevo sends transactional email from the EU, Crisp powers support chat from the EU and the United States, Google Analytics handles analytics in the United States and Whop processes payments in the United States. The privacy policy acknowledges international transfers, notably to the United States, under Standard Contractual Clauses or adequacy decisions. Two things are not offered: any choice of hosting region for the customer, and any hosting certification. DNS resolves behind Cloudflare, on the anycast address 172.67.75.176.
Things to keep in mind
Risks and trade-offs to weigh before adopting Pixel Dojo.
- No GDPR commitment appears anywhere on the site, no DPA and no EU representative, and the exclusion from model training is conditioned on explicit consent whose procedure is nowhere documented and offers no opt-out
- The existence of a free tier is contradictory: the homepage JSON-LD announces a free tier with included credits while the pricing page shows only paid plans, so do not assume one exists before checking at sign-up
- Session images vanish after one hour unless you click Save or Download, so a closed tab is enough to lose the work you just paid credits for
- Credits are unforgiving: once consumed they are never refunded and carry no monetary value, dated promotions of up to 40% can move per-model pricing, and a chargeback triggers immediate suspension of access
- Model counts differ from page to page, 70+ on the homepage, 125 tools on the pricing page and 153 models on the API page, and because the models are third-party their availability can change
- The publishing entity is hard to pin down: BLOVE INC appears only in the privacy policy while the terms and footer show the trade name alone, and the domain was registered on 08/02/2024, after the first Wayback capture of 21/12/2021, leaving the domain history uncertain
- Moderation is strict and largely automated, so prompts and images can be blocked including wrongly, with a human review offered on request; more broadly, a pipeline this fast can quietly erode your own craft judgement, so review outputs critically, verify third-party rights before publishing, and remember that a convincing image is not a lawful or accurate one
Setup & Integrations
Technical difficulty
Low for everyday use. Create an account, open a tool, write a plain-language prompt and keep the defaults; the FAQ states that no design or technical experience is required. The real friction is the credit system and the variable cost of each model. The agent route is the easiest technical path: one terminal command, then an approval in the browser, with no API key to copy. The API route is developer-level: create a key, buy credits and make asynchronous REST calls with a job ID and a status URL. Zapier and ComfyUI integrations cover existing workflows.
Deployment
Integrations
Behind Pixel Dojo
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Pixel Dojo.
Frequently asked questions
How many AI models does Pixel Dojo give access to?
How much does Pixel Dojo cost?
How do the credits work?
Can I use the images and videos commercially?
Can I cancel, and is there a refund?
Do I need design or coding skills?
Are my generations private, and do they train the models?
Is there an API, and can an AI agent drive it?
Is there a mobile app, and is NSFW content allowed?
Who operates the service and where is the data hosted?
Should you pick Pixel Dojo?
Pixel Dojo does not build models, it aggregates them. That single choice defines both its value and its limits. What you buy is breadth at one price: image, video, audio, upscaling, editing, LoRA training and three job-level Studios behind a subscription that starts at $10 a month. For a creator who was paying separately for Midjourney, Runway, Magnific, Luma and Leonardo, the value for money is genuinely defensible, and the publisher's own $164 against $25 comparison is the honest expression of the pitch.
Operationally the platform looks solid. Safety is documented layer by layer, subprocessors are named one by one with purpose, data types and location, session images are not persisted by default, credits are charged only on success, and a changelog has been kept since April 2025. The developer surface, a public REST API in every plan and a hosted MCP server usable from Claude Code, Cursor or Windsurf, is more than most creative tools offer.
The weak spot is the compliance file. GDPR is never mentioned, no DPA is offered, no Article 27 representative is appointed and no certification is claimed. The publishing entity is equally discreet: BLOVE INC appears only in the privacy policy, there is no about page, no contact page and no named team, and one email address handles support, legal, privacy and abuse. Add a dependence on third-party providers whose availability and credit pricing can shift, and an exclusion from training conditioned on explicit consent with no documented mechanism.
The verdict follows naturally. Pixel Dojo suits individual creative production and small teams that value breadth, speed and a predictable bill. It suits a regulated enterprise purchase far less, where the missing contractual guarantees will end the conversation before the models ever do.
- Choosing a selection results in a full page refresh.
- Opens in a new window.