Relay services and the move to real-time text — library

Relay services: what the network owes

Telecommunications Relay Service (TRS) is the set of services that let people who are deaf, hard of hearing, deafblind, or have a speech disability use the phone network on functionally equal terms. It is mandated under Title IV of the Americans with Disabilities Act, and the internet-based forms are compensated from the Interstate TRS Fund, which the FCC oversees. The main forms:

VRS (Video Relay Service) — a user signs to a video interpreter (a communications assistant), who voices to the hearing party and signs the reply back. Requires broadband video.
IP-CTS (IP Captioned Telephone Service) — a user speaks and listens normally but reads real-time captions of the other party’s speech, delivered over an IP connection.
IP Relay — text-based relay over the internet.
Traditional (TTY-based) TRS and Speech-to-Speech — the older text-relay and speech-assistance forms.

These are not niche add-ons; they are federally mandated capabilities of the phone system, and the fund that pays for them is part of the same regulatory apparatus as universal service.

How the TRS Fund works

The Interstate TRS Fund is financed by mandatory contributions from telecommunications and VoIP providers, assessed as a percentage of end-user revenue — a “contribution factor” set annually by the FCC’s Consumer and Governmental Affairs Bureau, structurally parallel to the universal-service contribution. There are actually two factors: one for the older non-internet-based services (assessed on interstate and international revenue) and a higher one for internet-based TRS — VRS, IP-CTS, and IP Relay — assessed on intrastate revenue as well.

The fund is administered under FCC contract by Rolka Loube, which each spring recommends the compensation formulas, budget, and contribution factors that the Bureau then adopts. For the 2025-26 fund year the net funding requirement was about $1.5 billion — and the cost is dominated by the two internet-based services, VRS and IP-CTS, which together account for the large majority of payments. Because the factors reset every summer, the specific percentages are a point-in-time figure, but the shape is stable: a billion-and-a-half-dollar obligation carried on the same shrinking telecom-revenue base as USF.

How VRS providers are paid

VRS is the most expensive and most closely regulated of the relay services, and its economics are worth understanding because they show how the fund actually sets prices. Providers are compensated per minute on a tiered schedule — larger providers earn a lower marginal rate as their volume rises — under a five-year plan the FCC adopted in 2023 with annual inflation adjustments tied to a labor-cost index. For the 2025-26 fund year the rates ran from about $8.33 per minute for the smallest providers down through roughly $6.73 and $4.21 per minute for the higher-volume tiers, plus a small additive for video-text service.

The FCC has also reshaped how VRS may be provided. It authorized skills-based routing (directing legal, medical, or technical calls to specialist interpreters) and the use of Certified Deaf Interpreters, and — significant for the workforce — it raised the share of a provider’s minutes that may be handled by at-home interpreters from half to 80 percent while easing the experience requirement to work from home. These are the kinds of operational rules that determine whether the service can actually scale to meet demand.

The TTY-to-RTT transition

TTY (teletypewriter) is the legacy technology for text communication over the phone line — decades old, and built for the analog, circuit-switched network. The problem: TTY tones do not survive IP networks reliably. The Baudot tones TTY uses were designed for analog circuits, and modern voice codecs, jitter, and packet loss corrupt them, so as the network moves to VoIP and all-IP, TTY becomes unreliable exactly where it is most needed, including for 911.

Real-time text (RTT) is the IP-native replacement. RTT transmits text character by character as it is typed, so the other party sees the message forming in real time — closer to a spoken conversation than a send-and-wait text — and it rides inside the IP media session alongside voice, which means it is native to IMS/VoLTE rather than bolted onto it. It is interoperable across networks and devices, backward-compatible with TTYs during the transition, and required to support 911 and 711 (relay access).

The FCC adopted RTT rules in its Report and Order FCC 16-169 (December 2016), which permitted wireless carriers to support RTT in place of TTY as they moved to IP networks and set a phased schedule — the nationwide carriers first (support beginning at the end of 2017, extending across new devices through 2019), regional carriers following into 2020–2021. RTT is now broadly deployed as a native handset feature by the major nationwide carriers on their VoLTE networks.

One nuance worth keeping straight: TTY has not been formally sunset. Backward compatibility with TTY is still required so the two user populations can reach each other, and the obligation to support TTY where technically feasible persists — particularly on legacy wireline networks that cannot carry RTT. What changed is that carriers may stop supporting legacy TTY over their own IP networks once RTT is in place, and several have received waivers to do so, conditioned on consumer notice. There has been little Commission-level RTT action since the early 2020s, so the framework in force is essentially the 2016 order as implemented.

Relay and caller authentication

Relay calls sit awkwardly inside STIR/SHAKEN, and the FCC put the question on the record: paragraphs 123 and 124 of the KYUP FNPRM (FCC 26-32) asked how the STIR/SHAKEN requirements apply to TRS providers, and whether TRS providers should be folded into a revised definition of “voice service provider.” Four providers answered on August 10, 2026 — Sorenson/CaptionCall, Hamilton Relay, InnoCaption and ZP Better Together — and they agree on the diagnosis.

The mechanical problem is credentials. InnoCaption states that TRS providers “cannot obtain Service Provider Code (‘SPC’) tokens, as they do not meet the STIR/SHAKEN Governance Authority’s requirements to obtain a token,” and that many lack control over the underlying network. Without an SPC token there is no certificate, and without a certificate there is no signing.

The second problem is the shape of the call. Hamilton Relay draws the line by service type: for IP CTS the current framework “generally reaches the right result, so long as the user’s voice service provider assigns the attestation the call has earned,” but for PSTN-based relay it does not, because “the conferenced nature of these calls means they cannot satisfy the criteria for A-level attestation.” A relay call is a three-party construction, and A-level attestation was written for a provider vouching for its own originating subscriber.

The consequence lands on users. ZP Better Together quoted the Notice back at the Commission: downstream filtering of unsigned VRS calls “disproportionately harm[s] individuals with disabilities.” Its ask is a mandate that TRS calls receive A-level attestation from their underlying providers, grounded in §225(b)(1)’s requirement that relay be provided in the most efficient manner.

Sorenson and CaptionCall filed the structural version. VRS and IP CTS providers “perform the functions of an initiating provider,” and A-level attestation “consists of three components that can be performed by different providers” — so the Commission could require originating providers to recognize a trusted certificate carrying the TRS provider’s attestation of end-user verification, rather than forcing the relay provider into a role it cannot occupy. The alternative Sorenson offers is expanding SPC token eligibility to internet-based TRS providers, with an interim mandate of A-level attestation for authenticated relay calls. It also argues the existing VRS and IP CTS user-registration and identity-verification regimes already satisfy KYC and KYUP.

The mechanism most of these filings reach for is a delegate certificate — a credential carrying an attestation made by a party other than the one holding the SPC token. Reply comments are due September 8, 2026.

Why it matters for the transition

Accessibility is the consumer-protection obligation most easily lost in a network migration, because it is the one with the least commercial constituency behind it. TTY breaking over IP is a concrete example of what “the transition strips things at the seams” means for real users. The design principle that follows — build real-time text, relay, and multimedia as native properties of the all-IP network rather than retrofits — is the same one that applies to caller identity and rich call data: features carried by the network end to end, not accessories grafted on afterward.