Focal
Focal is a browser-based AI video studio that turns an idea or a script into finished TV episodes, films and short-form clips, generating shots, voice-overs and sound effects, then polishing them on a conventional timeline.
What is Focal?
Focal is a browser-based video production tool from Fluorescent Labs, a two-person San Francisco company out of Y Combinator's Winter 2024 batch. Its homepage calls it the agentic video creation platform, and the ambition it states is larger than a single clip: making your own TV shows and movies.
The workflow runs end to end. You either bring your own script, which Focal adapts faithfully, or start from a one-line idea the tool expands into a script for you. From there the script stays editable in conversation — the site's own examples are instructions such as make this conversation shorter, or asking for a series of over-the-shoulder shots aimed at whoever is speaking. Focal then generates the shots, the images behind them, the voice-overs and the sound effects, and hands the result to a conventional timeline editor for trimming and polishing, with video extension and frame interpolation available at that stage. Characters and locations can be generated once and reused, which is what makes a series possible rather than a set of disconnected clips.
Rather than train its own models, Focal aggregates other people's. As of August 2026 it draws on Veo, Seedance, Kling and Hailuo for video, GPT Image, Nano Banana (Pro), Z-Image, Flux Klein and Seedream for images, and ElevenLabs, InWorld and OpenAI for voices, with ElevenLabs also handling sound effects.
That choice is exposed to the user as three documented quality modes. Fast costs roughly 18 credits per minute of video, runs on lighter models, is the only mode available on the free plan, and openly admits to inconsistent motion and characters. Pro sits at around 190 credits per minute on Seedance 1.5. Max, at roughly 600, pairs Seedance 2.0 with Opus 4.6 for the most demanding work.
Access is web-only: no mobile app, no desktop client, no browser extension and no public API, though the FAQ says one is being worked on. Around the product sit a blog of viral-format tutorials, a YouTube channel and a Discord community.
What it does
- Turn an idea or an existing script into a finished video
- Generate video shots, images, AI voice-overs and sound effects
- Rewrite the script through plain-language chat commands
- Create characters and locations once and reuse them across projects
- Cut and polish the result on a conventional timeline
- Trade speed against quality through three generation modes
- Produce vertical formats for TikTok and YouTube Shorts
When to use Focal / When not to
A quick filter to help you decide if Focal is the right fit.
When to use Focal
- Solo creators and indie filmmakers who want to produce narrative video without an editing team
- Marketers and brand teams turning out explainer videos, ads and social content at volume
- Educators and instructional designers who need lessons rendered as watchable video
- Short-form creators working in vertical formats for TikTok and YouTube Shorts
- Writers adapting an existing script — or a book — into moving images
When not to use Focal
- Developers who need to drive the tool programmatically: there is no public API yet
- Anyone looking for transcription, video summarisation or upscaling, all of which the FAQ says are absent
- Organisations that require a DPA, a published subprocessor list or a SOC 2 / ISO certification
- Users unwilling to have their prompts and outputs train the models, since no opt-out is documented
- People wanting to work commercially on the free plan, or from a mobile or desktop app
How to use Focal
A typical end-to-end flow, from setup to results.
- Create a free account from the Get Started Free button — registration is required to use the service
- Describe your idea in the prompt box on the homepage, or import a script you have already written
- Pick the language and the input mode; the models shown by default are Auto, GPT Image 1.5 and Seedance 1.5 Pro
- Let Focal draft the script, then refine it with plain-language chat commands
- Generate the characters and locations your story needs, so they stay consistent from shot to shot
- Generate the shots themselves, along with the voice-overs and sound effects
- Adjust the result on the timeline, extending clips or interpolating frames where needed
- Switch between Fast, Pro and Max depending on how many credits you are willing to spend per minute
- Keep an eye on the monthly and daily credit balance, which is what actually limits output
- Reach the team through the in-app Support button once logged in; manage or cancel the subscription from Profile or Pricing
Pros & Cons
Pros
- The whole chain — script, shots, voice, sound, edit — lives in a single browser tab
- Access to competing state-of-the-art models rather than one vendor's output
- Characters and locations stay consistent from one project to the next
- Three cost tiers with the credit cost per minute of video published openly
- A permanent free plan, and a paid entry point at 10 USD per month
- Commercial use is unlocked from the cheapest paid plan onward
- The company claims no ownership of inputs or outputs and assigns output rights back to the user
Cons
- No public API, so nothing can be automated from outside the product
- The free plan carries zero starting credits and no monthly or daily allowance
- Inputs and outputs are used to train the models, with no documented way to opt out
- The licence granted over your content is broad, sublicensable and irrevocable while it is stored
- Subscriptions are non-refundable and no partial credit is given
- No DPA, no published subprocessor list and no security certification for buyers who need them
- The about page is out of date, still naming Minimax, Luma and Runway where the FAQ lists Veo, Seedance, Kling and Hailuo
Pricing & Plans
Focal operates on a freemium basis. A permanent free plan is available at no cost, restricted to the Fast generation mode, to basic models and to personal use, and carrying no starting, monthly or daily credits. The cheapest paid plan, Starter, is priced at 10.00 USD per month and is the point at which commercial use, the full range of models and a monthly credit allowance become available. All plans are billed monthly in US dollars; no annual option is offered, and subscriptions are non-refundable.
- no starting credits
- no monthly or daily credits
- basic image and video models only
- slower generations
- personal use only
- 1
- 000 monthly credits
- 50 daily credits
- access to the best models including Veo and Seedance
- up to 1
- 000 seconds of video per month
- commercial use
- 3
- 900 monthly credits
- 50 daily credits
- access to the best models
- up to 3
- 900 seconds of video per month
- commercial use
- 15
- 000 monthly credits
- 50 daily credits
- access to the best models
- up to 15
- 000 seconds of video per month
- commercial use
Data, GDPR & hosting
A consolidated view of how Focal handles your data.
GDPR overview
The privacy policy, last updated 29 January 2025, carries a dedicated section on user rights under the GDPR: withdrawal of consent, objection, access and copy, rectification, restriction, erasure, portability and the right to lodge a complaint with a supervisory authority. Legal bases are enumerated, requests are free and answered within one month, and objection to direct marketing is unconditional. Fluorescent Labs is named as data controller, reachable at machines@focalml.com. Beyond that the implementation is thin: no Article 27 EU representative is designated, no data protection officer is named, no DPA is published, no transfer mechanism such as standard contractual clauses is cited, and the subprocessor list is not published — it is only available on request. The company states it is US-based and that data may travel to jurisdictions offering weaker protection.
Who owns the data?
Focal's terms state plainly that the company claims no ownership of your inputs or your outputs, and that any rights it might acquire in your outputs are assigned back to you. In exchange you grant Fluorescent Labs a worldwide, royalty-free, transferable and sublicensable licence — irrevocable for as long as your content is stored with them — to host, reproduce, distribute, display and process that content in order to operate, improve and promote the service, including sharing it with third parties under contract. Commercial exploitation of your outputs is permitted only on a paid plan; the free plan is for personal use only.
Reuse rights
On a paid plan, whatever you generate is yours to reuse, publish or sell without asking permission, and the terms expressly assign back any rights the company might otherwise hold in your outputs. Two limits bound that freedom. On the free plan, outputs are for personal use only and may not be monetised. And because generative models are involved, Focal warns that your outputs may not be unique: other users can produce identical or substantially similar results, and you waive any claim arising from that overlap. Separately, the company retains its own broad licence over everything you upload or generate, including the right to use inputs, outputs and interactions to train and improve its machine-learning models, with no opt-out documented. You remain responsible for holding the rights to whatever you feed in.
Data retention & training
Hosting summary
Focal publishes no hosting information in the usual sense: no provider is named, no region can be chosen and no data-residency option is offered. What the privacy policy does say is that Fluorescent Labs is based in the United States, that processing takes place at the company's operating offices and wherever the parties involved happen to be, and that it may use service providers operating in other countries. Personal data may therefore be transferred to the United States or to other jurisdictions whose data protection laws offer weaker safeguards than the user's own. No transfer mechanism is cited — standard contractual clauses are not mentioned — and the subprocessor list is not published, only available on request. The practical reading is that data sits under United States jurisdiction, with San Francisco as the company's stated base. The site itself is served through Cloudflare's anycast network, but that is content delivery infrastructure and says nothing about where projects, prompts or generated videos are actually stored.
Things to keep in mind
Risks and trade-offs to weigh before adopting Focal.
- Everything you type and everything you generate feeds model training, under a perpetual licence with no opt-out on offer
- The licence you grant over stored content is worldwide, sublicensable and irrevocable for as long as the content sits on their servers
- Focal is under no obligation to keep your work or give you copies of it, so export anything you care about
- Using the voice or face features means handing over biometric data, kept for up to three years after you delete your account
- Personal data feeds interest-based advertising and is shared with third-party advertising partners
- Subscriptions renew automatically each month and are non-refundable, so cancelling has to happen before the renewal date
- Output comes from a generative system: Focal aims at a PG-13 level without guaranteeing it, and near-identical results may surface in someone else's project
Setup & Integrations
Technical difficulty
Very low. Focal is a web application: nothing to install, no API key, no integration to configure. An account is required, after which the entry point is a prompt box you write into in plain language, and the about page claims no complex editing skills are needed. The one thing genuinely worth understanding before starting is the credit system: each plan carries a monthly and a daily allowance, and the Fast, Pro and Max modes consume roughly 18, 190 and 600 credits per minute of video respectively.
Deployment
Integrations
Supported languages
Behind Focal
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Which AI models does Focal use?
Can I keep the same characters and locations across a project?
Do I own the videos I create, and may I sell them?
Is my content used to train Focal's models?
What does Focal cost?
Is there an API?
Is there a mobile app?
Can Focal transcribe or summarise videos?
Can I get a refund if I change my mind?
Where is my data processed?
Should you pick Focal?
Focal is not a text-to-video toy. It is an attempt at a full production chain — script, shots, voices, sound effects, timeline — assembled in a browser and aimed at episodes and films rather than fifteen-second clips. Two decisions make it interesting: aggregation and consistency. By wiring in Veo, Seedance, Kling, ElevenLabs and half a dozen image models instead of training its own, Focal lets its users follow the state of the art rather than being pinned to one vendor's quality curve. By making characters and locations reusable, it makes serialised work possible at all. The three quality modes, with their credit cost per minute published openly, are an unusually honest way of exposing the speed-quality-price trade-off, down to admitting that Fast mode produces inconsistent motion.
The reservations are contractual rather than technical. Inputs and outputs feed model training under a perpetual, sublicensable licence with no opt-out documented; the licence granted over stored content is broad and irrevocable while it lasts; and nothing obliges the company to keep or return your work. Buyers with procurement requirements will find nothing to hand them — no DPA, no subprocessor list, no certification, no EU Article 27 representative. Subscriptions do not refund.
Set against that, the entry price is low and honest: 10 USD per month unlocks the full model range and commercial rights, and the free plan is a genuine look rather than a countdown. Behind it is a two-person company from Y Combinator's Winter 2024 batch, shipping visibly — the homepage was advertising Seedance 2.0 in August 2026. Small team, fast movement, permissive output rights, demanding data terms: worth trying for creative work, worth reading closely before anything sensitive.
- Choosing a selection results in a full page refresh.
- Opens in a new window.