Product & Market
The Signet Captioning API, customer segments, pricing baseline, market sizing, and competitive landscape.
First product — the Signet Captioning API
Input: ASL video. Output: English translation suitable for asynchronous captioning workflows, timestamped text in the production layer, and optionally the pose representation for downstream use.
The commercial API is operated by the Signet project, not by miners. Miners compete to produce the artifact. The project deploys the champion.
The product enters asynchronous or reviewable workflows first, then expands toward more demanding conversational use as measured accuracy allows.
Who buys it
The economic buyer is an institution, not the deaf user. Accessibility is purchased by organizations with legal, contractual, procurement, or audience-access obligations.
Initial segments:
- educational content providers
- broadcasters and media archives
- accessibility technology vendors
- government and public-service media
- corporate training libraries
- transit and public-information systems
Regulatory demand is active under the European Accessibility Act, the Americans with Disabilities Act, Section 508, and related obligations.
Pricing baseline
| Tier | Price | Terms |
|---|---|---|
| Pay-as-you-go | $0.15 / minute | No commitment |
| Volume | $0.09 / minute | >50K minutes/month, annual contract |
| Enterprise | $0.06 / minute | >500K minutes/month, SLA |
| Compliance subscription | $2,400 / month | Includes 20K minutes and reporting |
The price sits materially below manual interpretation, and is designed for content volumes where human interpretation is uneconomic.
Planning baseline, August 2026.
Market size, stated honestly
The August 2026 market baseline records syndicated estimates of roughly $1.2B–$1.8B for the broader sign-language translation and interpretation categories, with higher projected values later in the decade.
Those figures include human services and adjacent products, so they overstate the immediately addressable automated software market.
Signet treats the real target as an emerging low-hundreds-of-millions automated software category, not the entire interpretation economy.
Competitive landscape
The field contains large human-service incumbents and a thin group of AI-native companies. The project research baseline identifies Signapse, SignAll, Hand Talk, KinTrans, Signvrse/Terp 360, EQ4ALL, SiMAX, and major platform research from Google and others.
The gap: there is no dominant open, continuously improved, private-benchmark-selected sign-language translation model that outside developers can deploy as a transparent artifact.
Position inside Bittensor
The audited SN1–SN128 landscape contains no subnet whose core task is sign-language translation.
The nearest semantic neighbor is speech interpretation, especially SN59 Babelbit. Audio interpretation does not solve visual linguistic understanding, pose extraction, or sign-language grammar.
Signet occupies a distinct model-R&D category inside Bittensor.