Session vs Deepgram
Side-by-side trajectory, velocity, and editorial themes.
Session shipped a protocol rewrite and a paid tier, then went publicly broke — the founder is asking users to bail it out.
Session is simultaneously in its most ambitious technical phase and an open funding crisis. Protocol V2 — re-implementing forward secrecy and layering post-quantum cryptography on top of Session's onion-routed transport — has been announced, and the Session Pro paid tier exited beta planning into a December development update. Then in March, cofounder Chris McCabe published a personal appeal saying the project cannot continue developing without user support, and the public feed has been quiet since.
The product roadmap that was meant to fund itself via Session Pro is colliding with the underlying problem the appeal makes plain: the Loki/Oxen-era token economics and donations aren't covering ongoing development. Protocol V2 and Pro are the bets that have to land for Session to remain viable; if Pro doesn't convert a meaningful share of the user base, the next twelve months are about scope reduction, not feature growth. The Feb 1 APT key rotation in January suggests the core infrastructure is still being maintained — for now.
Watch for either a hard Session Pro launch and conversion announcement, or a more explicit wind-down / handoff post. A long stretch of silence after a funding appeal usually resolves one way or the other within a quarter; the absence of any new posts since mid-March is itself a signal.
Diarization v2 lands with a 3.3× human-eval edge — Deepgram's contact-center push gets sharper.
Deepgram is shipping in two coordinated lanes: deeper transcription quality (Nova-3 multilingual numerals, Gujarati, profanity filtering across 50+ languages) and a maturing Voice Agent API (managed LLM swaps, third-party TTS controls). The new opt-in diarize_model=v2 brings a new architecture preferred 3.3× over v1 in human eval, with the biggest gains on contact-center audio. Self-hosted images and multi-language SDKs are released on a tight, predictable cadence.
The arc is consolidating around enterprise contact-center workloads: better speaker separation, safer outputs via profanity redaction, and richer language coverage are exactly the gates that block call-center adoption. Voice Agent is becoming a managed-LLM thin layer where customers pick the brain (OpenAI, removed Llama Nemotron) while Deepgram owns ears and mouth. Expect diarize_model=v2 to become the default once telemetry catches up.
Likely next: v2 diarization promoted to default for diarize=true, and a streaming version of the same architecture to extend the contact-center story to live transcription. More managed-LLM additions in Voice Agent, plus continued language fill-in for Nova-3.
See more alternatives to Session →
See more alternatives to Deepgram →