Encrypt all your AI chats.
One OpenAI-compatible API to hundreds of models, with a privacy label on every route: zero retention, end-to-end encrypted, EU.
Moving from OpenRouter is one base_url change. Start with 1M free tokens on DeepSeek, no card:
https://t.co/Ne0VXd2xOo
Enterprises buying AI through hyperscalers overpay by up to 10x, and have to trust the host with every prompt.
@TrustedRouter just raised $6M, led by @Sierra_Ventures, to fix both with the NY Token Exchange @nytokenexchange
Buying AI for a company?
On the New York Token Exchange you set the privacy minimums, and approved providers compete on price within your spend limits. Your prompts stay confidential.
Request an allocation: https://t.co/I0CO7RcWoE
BREAKING 🚨 @TrustedRouter raises $6M seed led by @Sierra_Ventures for its private gateway to hundreds of AI models, which offers verifiable confidential computing on some open-weight models, according to Semafor
https://t.co/ybsm3hURwe
Only 16 of the 98 providers on @TrustedRouter clear its zero-data-retention bar on every route.
Telnyx is one of them.
TrustedRouter is an OpenAI-compatible AI gateway that says it keeps no prompt or output logs.
All 17 of our hosted models are now listed there, running on GPUs we own.
Our ZDR is the commitment in the Telnyx privacy policy.
Chat completion requests are processed in memory and discarded once the response returns, and we keep request metadata like timestamps and token counts.
TrustedRouter's listing labels it policy-backed ZDR.
You can checkout Telnyx TrustedRouter listing at https://t.co/TwQGUxL6PW and our retention terms at https://t.co/GFhgzrt6Fl
i made `mev` and `trev` models (including `jev` available on TrustedRouter) that give structured output; but also you can enable reasoning if you want a decision but with more correctness
mev uses the excellent, fast, american Mercury model