For example, a storyboard artist needs to produce four shot frames for the third scene he wrote himself. In Laper, he cannot open a blank input box and type a prompt; he first has to go to this scene, fill in the shot size and camera position in the shot panel, and then initiate generation. When the result comes back, the image is attached to this project and becomes an auditable asset. This is the workflow described in Laper's official product article (AI Production Assets Grounded in Project Entities[1]).
This rule—first have a project entity, then generate—is the most obvious divergence between Laper and tools that put generation in a blank prompt box. Laper calls itself an all-in-one creative OS for screenwriters, directors, and storyboard artists, analogous to Cursor, CapCut, and Figma in visual storytelling (TrustMRR project page[2]).
The TrustMRR page updated on 2026-11-14 02:37 shows that Laper's MRR is $52,789, all-time revenue is $219,476, and it has 1,641 active subscriptions; the page says the revenue is verified through a Stripe API key (TrustMRR page[2]). MRR is monthly recurring revenue, not profit; this material contains no costs, gross profit, or customer acquisition spending, so whether it is making money remains unclear.
Image: Laper official display[3], accessed 2026-11-10, original interface preserved.
After Generation, Where Does the Image Go?
What is really worth examining is where the image lands after generation and whom to reconcile with when something goes wrong. Users initiate tasks from a character, scene, or shot; Laper's backend handles deducting credits, dispatching, and determining success or failure; successful outputs become project assets and, when needed, are attached back to entities such as characters and scenes through the normal path, and one task can produce multiple images. Independent generation services that call models and upload images cannot verify who you are, nor can they directly modify your project (product article[1]).
In other words, AI has no authority to declare “this job succeeded.” On success, the backend completes the task and writes the asset into the project; on terminal failure, “the applicable refund policy may apply.” Users receive task notifications, but the final reconciliation looks at the persistent records in the project.
What is easy to miss is where the image ultimately attaches: AI-generated images are not saved separately in a hidden document set, and when needed they return under entities such as characters and scenes. Character profiles request portraits or casting references, locations request visual references, props request design references, scenes request stills based on screenplay scope and characters, and storyboard shots request frames using scene text plus explicit shot settings. The Assets module is not a general-purpose generate button; it aggregates the generation canvas and task list; generation must be initiated from a character, scene, location, prop, shot, or tool panel.
This sacrifices the ease of “just generate one and take a look,” in exchange for every image being able to answer: who initiated it, and which project it belongs to. In multiplayer collaboration, the latter is worth more.
Two People Editing the Same Passage
When two people edit the same passage of a screenplay, Laper's official statement is: a paragraph is the conflict boundary, and the later-resolved value wins. This is node-level collaboration, not character-by-character merging within a paragraph (Real-Time Screenplay Collaboration With Loro CRDT[4]). In reality, if one person edits finely and another overwrites the whole paragraph, the fine edits lose.
Opening an existing project has another gate: it must first sync online. The client must wait for the server snapshot, subsequent updates, and an explicit sync-complete signal before the editor becomes editable; the local cache is only for caching and recovery and cannot bypass this server-first gate. If the network briefly drops after opening, local operations can enter the persistence pipeline, and sync is caught up after reconnecting.
Pricing and What It Doesn't Say
The pricing page has five tiers, from free to $400/month (Laper Pricing[5]). Most relevant to the main thread are credits: the Senior tier costs $20 per month and gives 600 credits for image and video generation, plus 20 refresh credits per day. Annual payment says “save more,” but no actual paid price is listed.
The credits terms are not generous to users: subscription credits are valid only during the subscription term; if unused, they expire at the end of each billing period and do not roll over, and they are immediately forfeited upon cancellation or non-renewal; directly purchased credits do not expire as long as the account remains active and in good standing. All credit purchases are final and non-refundable; subscriptions do not offer partial-period refunds, cancellation takes effect at the end of the current billing period, and refunds are not prorated; the only exceptions are duplicate charges due to technical errors on Laper's side, suspicious transactions reported within 48 hours, and statutory refunds, and these are at Laper's sole discretion (Credits Policy[6], last updated 2025-12-08).
This set of rules is friendly to cash flow and not exactly lenient to users. The product article says that upon terminal failure, “the applicable refund policy may be applied,” but the terms only state that credit purchases are non-refundable, subscriptions do not receive partial-period refunds, and there are a few exceptions; whether AI task failures refund credits or cash, and how much, is not written in the terms.
Two Situations Where It Says It Should Not Be Chosen
Interestingly, Laper itself spells out the scenarios where it does not fit: if the need is a traditional file-centric production workflow, Final Draft may be the more familiar center; if priority is real-time collaborative writing, video calls, and broad device support, WriterDuet is a mature collaboration-first product (Real-Time Screenplay Collaboration With Loro CRDT[4]). These alternative recommendations come from Laper's own product article, not an independent comparison review; it is itself rare for someone selling software to proactively point out two situations in which they should not be chosen.
At the top of the same TrustMRR page, this startup is for sale, asking $1,999,999, with $1,470,500 displayed next to it and labeled 2.2x revenue. This is a listing, neither a transaction price nor a financing record.
Back to That Storyboard Frame
AI generation itself is no longer difficult; what is difficult is the order after generation: which project an image belongs to, who handles a failure, whose edit wins when two people change the same passage, and how unused quota is calculated when a subscription is canceled. Laper writes ownership into the product structure and quota rules into the terms; what it saves is the effort of putting images back in place.