A doodle studio you chat with
Doodle AI is a free, browser-based AI studio that turns a photo — or a typed description — into hand-drawn doodle art: avatars, six-panel collages, sticker sheets, pet portraits, greeting images and printable line art. You steer it by chatting with an agent that picks the drawing style for you, or by opening a single-purpose tool page where the generator is already set up.
Doodle AI at a glance
| What it is | An AI image studio for hand-drawn doodle art, driven by a chat agent |
|---|---|
| What it generates | Still images only — avatars, collages, sticker sheets, portraits, greeting images, printable line art |
| Video generation | Not available (on the roadmap) |
| Input | A photo, a saved character, or a typed description |
| Drawing styles (skills) | 26 runnable skills across 4 categories |
| Images per generation | 1 for single-image skills, up to 9 for multi-image packs |
| Works without a photo | 2 of 26 skills |
| Price | Free. 1 credit per image, 10 starter credits on signup, no card |
| Account | Free Google sign-in required to create; browsing is open |
| Where it runs | In the browser, on Cloudflare's edge — nothing to install |
| Public gallery | None. Your work is account-scoped and private |
How the agent works
Most "AI photo" products are a grid of presets: you pick a filter and hope it is the one you meant. Doodle AI inverts that. You describe the outcome, and an agent decides how to get there.
Each skill is a self-contained package of instructions the agent reads before it draws, so the catalogue is the agent's actual capability list rather than a menu bolted onto a single model. On every turn the agent picks exactly one skill, calls a generation tool with your photo, and then offers three concrete next steps — a variation in the same style, a different skill that follows naturally, or a refinement pass. It can also read and rearrange your canvas, so a multi-image set arrives laid out instead of piled up.
Refinement is conversational. "Thicker outline", "warmer paper", "make the expressions more distinct" are all valid instructions — there is no prompt to hand-edit and no seed to manage.
What the skills can generate
Every drawing style is a skill. Both the chat and the tool pages draw on the same catalogue, so nothing is exclusive to one route.
| Category | Skills | Photo needed | Images per run | Examples |
|---|---|---|---|---|
| Avatars | 9 | Yes | 1 | Doodle Avatar, Chibi Mini-Me, Crayon Self |
| Collages | 2 | Yes | 1 | Doodle Collage, Full-Body Action Collage |
| Packs | 7 | Yes | 4 or 6 or 9 | Emotional Modes, Seasonal Pack, Expression Pack |
| Freeform | 8 | Some | 1 | Surprise Me, Sticker Pack, Mood Captions |
7 of the 26 skills are multi-image packs: one request returns a whole set — a run of expressions, four seasons, a sheet of stickers — rather than a single frame. The full list, with a real sample output for each, is on the skills page.
How Doodle AI compares
Three different kinds of product answer "turn my photo into art". They fail in different places.
| Dimension | Doodle AI | Preset filter apps | General-purpose image models |
|---|---|---|---|
| How you steer it | Describe the outcome; an agent picks the style | Pick from a fixed grid of filters | Write and tune your own prompt |
| Prompting skill needed | None | None | Substantial — the prompt is the product |
| Refining a result | Plain language, in the same conversation | Usually start over with another filter | Re-engineer the prompt, manage seeds |
| Multi-image sets from one request | Yes — up to 9 images | Rarely | Only by running it repeatedly |
| Reusing the same subject | Save a reference photo as a character, then @mention it | Re-upload each time | Re-attach or re-describe each time |
| Art direction | One consistent hand-drawn doodle house style | Whatever the filter set covers | Anything, which is also the problem — consistency is on you |
| Entry point for one specific job | 26 skills plus dedicated free tool pages | One app, many filters | A blank box |
| Cost model | Free; 1 credit per image, 10 free on signup | Often subscription or watermarked free tier | Per-image or subscription |
| Video | Not yet — images only | Sometimes | Increasingly yes |
How image generation works
Generation runs server-side through Doodle AI's own platform connection, so no provider credential ever reaches your browser and there is no API key for you to obtain. Chat is routed through OpenRouter; images are produced and hosted by PicX. Each returned image is a permanent URL you can download.
Credits are spent per image and cost the same whatever the style: 1 credit each, so a single avatar costs 1 and a 9-image pack costs 9. An image that fails is refunded automatically rather than silently charged. Credits are shared at the team level, not per person, so a workspace draws from one balance.
Roadmap
These are not available yet. They are listed because "what is planned" is a fair question, and because the honest answer is more useful than an implied one.
| Planned | Status |
|---|---|
| Video generation | Planned — not available. Doodle AI generates still images only |
| Checkout for credit packs | Planned — credits are granted manually during the testing phase |
| User-created custom skills | Planned |
| Character continuity across generations | Planned — saved characters exist today, identity-locking does not |
| WhatsApp and Discord access | Planned |
Frequently asked questions
What is Doodle AI?
Doodle AI is a free, browser-based AI studio that turns a photo — or a typed description — into hand-drawn doodle art: avatars, six-panel collages, sticker sheets, pet portraits, greeting images and printable line art. You steer it by chatting with an agent that picks the drawing style for you, or by opening a single-purpose tool page where the generator is already set up.
Is Doodle AI free?
Yes. Browsing is open to everyone, and creating requires a free Google sign-in. New accounts receive 10 starter credits, and no card is required. Each generated image costs 1 credit, so a single-image skill costs 1 credit and a multi-image pack costs 1 credit per image it returns. Images that fail are refunded automatically.
Do I need a photo to use Doodle AI?
Not always. 24 of the 26 skills work from a photo you attach, and 2 need no photo at all — you can describe a character in words and get a doodle back.
Does Doodle AI generate video?
No. Doodle AI generates still images only. Video generation is on the public roadmap but is not available today, and nothing on the site produces video.
Do I need any design or prompting skill?
No. The agent reads your request, chooses one of the 26 skills, generates the image, and then offers concrete follow-ups such as a different palette or a multi-image version. You can also refine in plain language — "thicker outline", "warmer paper" — instead of editing a prompt.
What this isn't
There is no feed and no public gallery. Your chats, boards, saved characters and generated results are account-scoped and visible only to you through the app. We do not build an advertising profile, and provider credentials are never exposed to the browser. It does not generate video, and it is not a general-purpose image model — it draws in one house style, on purpose.
Who made this
Doodle AI is built by Type-Think-AI, open source under the MIT licence, and deployed on Cloudflare's edge. The skill catalogue, the agent instructions and the generation pipeline are all in the public repository.
Turn a photo into a doodle
Free to start, 10 credits on signup, no card required.
Open Doodle AI →Or start from a specific job on the free tools page.