Ankon AI
Ankon AI turns a topic, script, PDF or reference image into a narrated whiteboard explainer video in minutes, writing the script, drawing each scene and burning in captions, with an SRT file and publish-ready metadata included.
What is Ankon AI?
Ankon AI is a generator of whiteboard explainer videos. You give it a starting point — a one-line topic, a script you have already written, a PDF, a text file or a reference image — and it returns a finished video in which ideas are drawn on a single white surface while a narrator speaks. The publisher calls the style ink-reveal: each scene appears stroke by stroke, timed to the voiceover, with no clip-art and no fade-ins.
Attached files become the subject rather than decoration. PDFs are extracted and images are described by a vision model, so the explainer stays faithful to what you supplied. Attachments are capped at three files and 3 MB combined.
Every finished job returns more than a video. You get an MP4 with optional burned-in captions, an SRT subtitle file that is always produced, a thumbnail, and a metadata file carrying an AI-written title, description, tags and chapter timestamps. Length runs from half a minute to ten minutes, two by default, in landscape 16:9 or vertical 9:16 for Shorts and Reels.
Narration covers nine language codes — American and British English, Spanish, French, Brazilian Portuguese, Italian, Hindi, Japanese and Chinese — with several voices each. A user may also register one cloned voice from a short recording, in English only; the clone runs on the publisher's own servers, requires explicit consent and a verified email, and is deletable on demand.
Pricing runs on credits: one credit is roughly one minute of finished video, with a two-credit minimum. Failed renders are refunded automatically, and a video delivered below 95% of the requested length credits the undelivered minutes back. Optional passes cost more — an independent fact-check adds one flat credit, a denser enhance mode doubles them, and a fast render adds 50%.
Beyond the web interface, Ankon AI exposes an asynchronous REST API and an MCP server, reachable with OAuth 2.1 from the Claude apps or with an API key from Cursor, Codex and other clients. A public community gallery lets anyone watch finished videos without an account.
What it does
- Write the narration from a topic, a script, a PDF or a reference image
- Draw whiteboard scenes stroke by stroke, in time with the voiceover
- Narrate in eight or more languages, with male and female voices
- Burn word-by-word captions into the video and export an SRT sidecar
- Write a publish-ready title, description and chapter timestamps
- Export landscape 16:9 or vertical 9:16 with a generated thumbnail
- Generate videos programmatically through a REST API or an MCP client
When to use Ankon AI / When not to
A quick filter to help you decide if Ankon AI is the right fit.
When to use Ankon AI
- Newsletter writers who want each issue turned into a short narrated clip for LinkedIn or TikTok
- Micro-SaaS founders who need a "How it works" explainer for a landing page without hiring an animator
- Course instructors converting written lecture notes into narrated, captioned modules
- Bloggers and business publishers sitting on deep written archives with no video presence
- Developers who want to generate videos programmatically through a REST API or an MCP client
When not to use Ankon AI
- Anyone who needs frame-by-frame control over every drawn stroke or over their own artwork
- Teams looking for an AI avatar presenting to camera rather than a drawn whiteboard
- Editors who want stock-footage montages cut automatically from an article
- Creators who need character animation or a full motion-design studio
- Non-English speakers hoping to narrate in their own cloned voice, which is English-only
How to use Ankon AI
A typical end-to-end flow, from setup to results.
- Sign up on the site and collect the five free credits, with no card required
- Choose your starting point: type a topic, paste a script, or upload a PDF or reference image
- Set the length, anywhere from half a minute to ten minutes
- Pick a narration language and a voice, or select your own cloned voice in English
- Choose landscape 16:9 or vertical 9:16, and decide whether captions are burned in
- Switch on the optional passes you need: fact-check, denser scenes, or fast render
- Launch the job — a one-minute short takes one to three minutes, a ten-minute video five to fifteen
- Download the MP4, the SRT file, the thumbnail and the ready-made title, description and chapters
- Developers: create an API key in the dashboard, POST to the jobs endpoint and poll until the status reads completed
- Or add the MCP server as a custom connector in Claude Code, Claude.ai or Claude Desktop and ask in plain language
Pros & Cons
Pros
- No editing software and no learning curve: a source goes in, a finished video comes out
- Captions, SRT, thumbnail and publishing metadata all arrive with the video
- Billed per delivered minute rather than per seat, with automatic refunds on failed or short renders
- Speech synthesis and voice cloning run on the publisher's own servers, not a third-party cloud service
- The publisher states plainly that customer inputs are never used to train any model
- A developer API and an MCP server are included and usable with the free credits
- A public gallery shows real output before you create an account
Cons
- Paid plans are not live: the upgrade page reads "Launching soon" and takes no payment
- No legal entity, postal address or named founder is published anywhere on the site
- The operation is barely a month old, the domain having been registered on 15 July 2026
- No frame-by-frame control over the drawing, and videos are capped at ten minutes
- Voice cloning is English-only, and videos made on free credits may carry a watermark
- Key-free OAuth works only from the Claude apps, and ChatGPT cannot connect at all
- Sub-processors are described by category with only the payment provider named, and no DPA is offered
Pricing & Plans
There is no permanent free plan. New accounts receive five free credits, presented as a USD 10 value and granted without a credit card for a limited period, which operates as a trial rather than a free tier. The lowest published paid entry point is the Starter plan at USD 23.00 per month, or USD 276.00 per year under the 20% annual discount. Prices exclude tax, which is calculated at checkout from the billing address. One credit corresponds to approximately one minute of finished video, subject to a two-credit minimum per video. It should be noted that, at the date of this review, billing had not yet opened: the plans page states that no payment is taken and no credits are granted until billing is live.
- Starter — USD 23.00 per month or USD 276 per year — 15 credits per month — for learning and personal projects
- Plus — USD 55.00 per month or USD 660 per year — 40 credits per month — marked Most Popular
- for YouTube creators
- Pro — USD 95.00 per month or USD 1140 per year — 75 credits per month — marked Best Value
- for agencies and professionals
- voice cloning highlighted
- Studio — USD 199.00 per month or USD 2388 per year — 160 credits per month — for teams and API workflows
- adding REST API
- MCP
- OAuth 2.1 and higher limits
- Premium Top-Up add-on — USD 2 per credit
- USD 10 minimum
- in tiers of 5
- 10
- 25
- 50 and 100 credits
- available with any active subscription
- Every paid plan includes 8+ languages
- no watermark
- burned-in captions
- SRT subtitles
- fact-checked scripts
- metadata and chapters
- fast render mode
- thumbnail download
- All four plans were shown behind a "Get early access" button and were not purchasable at the date of review
Data, GDPR & hosting
A consolidated view of how Ankon AI handles your data.
GDPR overview
The privacy policy, last updated in July 2026, addresses the GDPR directly without ever using the phrase "GDPR compliant". Readers in the EU or UK are told they may exercise access, rectification, erasure and portability rights and may complain to their local supervisory authority, with verified requests answered within 30 days; requests go to privacy@ankonai.com. For processing outside the user's country of residence, including in the United States, the publisher says it relies on Standard Contractual Clauses or equivalent safeguards with its sub-processors. A cloned voice is named as biometric data requiring explicit consent and a verified email. No Article 27 EU representative is designated, no data protection officer is named, and no Data Processing Agreement is offered anywhere on the site.
Who owns the data?
The terms are explicit: you own the videos you generate, to the extent permitted by law and by the upstream model providers' terms, and they are yours to publish anywhere. In return you grant Ankon AI a limited licence to process your inputs for the sole purpose of providing the service. You remain responsible for the material you submit and for the videos you produce with it. A cloned voice is treated separately as biometric data: the recording and the derived voice model stay on the publisher's own servers, are never shared with a third party, and can be deleted at any time, which removes both.
Reuse rights
The videos are yours to reuse without asking permission, and the publisher states they can be published anywhere. That covers the finished MP4, the SRT subtitle file, the thumbnail and the generated title, description and chapter timestamps. The limits are those set by law and by the upstream model providers' terms, plus the acceptable-use rules: nothing illegal or infringing, and a cloned voice may only be your own or one you have explicit permission to use, never a voice synthesised to impersonate or deceive. Videos stay private by default; listing one in the public community gallery is an explicit per-video choice.
Data retention & training
Hosting summary
Ankon AI names no hosting country and no hosting region. The privacy policy describes the pipeline by function instead. Narration is synthesised locally on the publisher's own servers using an open-source voice model, with no script text sent to a cloud text-to-speech provider, and a cloned voice stays on those same servers. Script and visual generation runs either on third-party AI model providers or on self-hosted models, with requests sent without project-attribution headers. Rendering happens on third-party cloud compute or the publisher's own render servers, and job data is held only for the duration of the render. Finished artifacts sit on third-party cloud storage and are delivered through time-limited links. The only geographic statement is that providers may process data outside your country of residence, including in the United States, under Standard Contractual Clauses or equivalent safeguards where the EU or UK GDPR applies. None of these providers is named except the payment processor. The domain itself is served behind Cloudflare on an anycast address, which reflects content delivery and says nothing about where data rests.
Things to keep in mind
Risks and trade-offs to weigh before adopting Ankon AI.
- The published prices are announced rather than chargeable: billing was not open at the date of review, so the grid may still change before launch
- No legal entity, postal address or named owner is published, leaving you with no identified counterparty if something goes wrong
- A hardware fingerprint is derived from your browser at signup and login, and your IP and browser type are logged for abuse review
- Free credits are capped at two accounts per device in any 30-day period, so a shared machine can quietly lose the offer
- Deleting your account forfeits purchased credits with no refund, and those credits are frozen whenever no subscription is active
- Narration written automatically from your source can sound authoritative while being wrong; the fact-check pass is optional and costs an extra credit
- A cloned voice is biometric data, and a video already shared to the public gallery stays there under an anonymised name even after you delete your account
Setup & Integrations
Technical difficulty
Very low through the interface: nothing to install, an email signup, and generation from a browser in three steps with no editing skills required. The API path is a conventional asynchronous REST flow — create a key, send it in a header, post a job and poll until it completes — with curl, Python and Node.js examples supplied. The MCP route is slightly fiddlier: you must type the client ID by hand in the Claude apps, since leaving it blank fails outright, and other clients need a key placed in a configuration file.
Deployment
Integrations
Supported languages
Behind Ankon AI
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Ankon AI.
Frequently asked questions
What can I make a video about?
How long does a render take?
Which languages are available?
Can I narrate in my own voice?
How does pricing work?
What happens if a render fails?
Are the videos watermarked?
Who owns the videos?
Is there a developer API?
Is there a minimum age?
Should you pick Ankon AI?
Ankon AI does one thing and names its own boundary: it turns written material into a narrated whiteboard explainer, and it says outright that anyone needing frame-by-frame control over the drawing should use a manual tool instead. Within that narrow scope the chain is complete. Narration, drawn scenes, voiceover, burned-in captions, an SRT sidecar, a thumbnail and a ready-made title, description and chapter list all arrive together, with no editing software involved.
Two things stand out for a tool this young. The first is the economics: billing is per delivered minute rather than per seat, failed renders are refunded automatically, and a video that comes up short credits the missing minutes back. The second is the developer surface, an asynchronous REST API and an OAuth 2.1 MCP server, both unusual at this stage and both usable with the free signup credits.
The reservations are about maturity more than about the product. At the date of review the paid plans were announced rather than sold: the upgrade page reads Launching soon and takes no payment, so the four published prices describe an intended grid rather than one in force. The site names no legal entity, no postal address and no founder; sub-processors are described by category with only the payment provider named; and there is neither a Data Processing Agreement nor a security certification. The domain was registered on 15 July 2026, which puts the whole operation under a month old.
Anyone sitting on a written archive with no video presence can test the claim at no cost, since five credits are granted without a card and the public gallery shows real output before signup. Anyone who needs an identified company, a contract or a purchasable subscription should wait for the next review.
- Choosing a selection results in a full page refresh.
- Opens in a new window.