AILANTA
← All research
GitHub · Hugging Face

Voice AI is scaling infrastructure before vertical workflows

The open market is rapidly expanding recognition, synthesis, and real-time models, but it has barely moved into specialized, regulated workflows. Verticalization remains a hypothesis rather than a validated mass shift.

17.07.2026368 GitHub919 Hugging FaceProduct Hunt excluded
Main finding

Verticalization has not happened yet

Specialized workflows represented 8% of the broad GitHub cohort in Q1 and 7.4% in Q2. Universal assistants and infrastructure remained above 92%. Explicit compliance features appeared in only 1% of Q1 projects and were absent from the Q2 cohort.

01

The universal layer remains dominant

Shares use only the broad voice AI, voice agent, speech agent, and voice assistant cohort.

Specialized workflows8% → 7.4%

1612

Universal solutions92% → 92.6%

185151

Compliance1% → 0%

20

Broad cohort201163

new projects

02

Healthcare is the only growing vertical pocket

Categories overlap; small absolute counts require cautious interpretation.

VerticalQ1Q2Change
Healthcare25+150%
Sales440%
Customer support31-66.7%
Scheduling42-50%
Finance20-100%

Healthcare rose from two to five projects, but none explicitly stated HIPAA, consent management, audit trails, or another compliance layer. This is early supply, not a mature regulated category.

03

Open-source infrastructure remains the core market

A layer must be explicitly present in a GitHub name, description, or topic.

LayerQ1Q2Direction
Speech recognition4546
Speech synthesis5640
Speech-to-speech1013
Telephony2316
Orchestration85

Speech recognition held at 45–46 projects, while speech-to-speech rose from 10 to 13. The decline in TTS and telephony cannot be separated from age bias because Q2 repositories had less time to reach three stars.

04

Hugging Face confirms infrastructure scale, not temporal growth

API cohorts are capped at the latest 1,000 models per filter, so this is a current snapshot of H1 models rather than a comparable Q1/Q2 time series.

Models919

624 with downloads

Publishers500

unique accounts

Vertical signals9.1%

84 models

Compliance2.7%

25 models

LayerArtifactsPublishers*Downloads
Speech recognition34018157,257
Speech synthesis7924341,070,061
Speech-to-speech6653
Telephony2215970
Orchestration16136,055

* Publishers are summed across quarterly cohorts and may repeat between Q1 and Q2. The full corpus contains 500 unique publishers. Vertical matches (29) are multi-label and do not represent the same number of independent products.

05

What true verticalization would require

01

Workflow must become more important than voice

The category will form when products are described through intake, claims, collections, or scheduling rather than STT, TTS, and latency.

02

Compliance must become a product feature

HIPAA, recording consent, auditability, data retention, and PII redaction must become architectural features rather than external promises.

03

Healthcare is the earliest candidate

It is the only vertical that grew in the GitHub cohort, but current supply still lacks the required regulatory layer.

Methodology and limitations

A public launch is defined as a new GitHub repository or Hugging Face model publication from January through June 2026. GitHub uses broad and vertical queries capped at two top-by-stars pages per query and month. Hugging Face is classified from metadata and README text. Product Hunt is excluded.

GitHub has an age bias from stars. Hugging Face top-by-recency cohorts do not provide a comparable time series; downloads are current snapshots. Workflow and compliance require explicit contextual evidence and are lower-bound estimates. Verticals overlap.