Overview
Making a film has never been cheaper. Getting permission to make one has never been harder: rights clearance, guild paperwork, distribution deals, festival calendars, insurance. The constraint on cinema has not been talent or tools for about fifteen years. It has been permission.
Longshot removes it. One continuous product takes an idea through development, screenplay, shot list and frames, and then releases the result to an audience in the same place it was made.
The bet
Two costs collapsed at once, and nobody had connected them:
- Production. A feature that cost $5–50M is now roughly $1,400 of inference.
- Distribution. Zero-egress object storage holds a feature for about nine cents a month and serves it to a million people for nothing. The bandwidth bill that killed every previous "YouTube for X" no longer exists.
Together: one person can make and release a feature for less than a laptop costs.
Two ideas worth understanding
No product code names a vendor. Every generative capability is a routing decision (modality, minimum quality, a hard cost ceiling, a preference) and the registry picks a provider, enforces the budget and fails over. Swapping model vendors is a config change, not a rewrite.
Money is a ledger, not a counter. Credits move through reserve → run → settle with compare-and-swap and idempotency keys, so unused holds refund and a failed generation refunds in full rather than silently charging for nothing.
Status
The pipeline is verified end to end in production on the free tier: signup through development, scene breakdown, an eight-shot list covering 38 seconds, and a 1024×576 still. Twelve credits, about a penny of retail value, with the ledger settling correctly throughout. The Studio and the API are live; the catalog schema exists and its interface does not yet.