|
Hey there 👋,
Quick refresher in case you're newer here: Dograh is an open source voice AI platform - build, test, and run phone agents on the model and telephony stack of your choice.
Last month was a big one: a v1.42.0 release packed with speech-to-speech models, self-serve credits that work across cloud and self-hosted, an MCP server that can now build voice agents for you, and a revamped UI - plus we're giving away credits for sharing your Dograh story.
|
A thank-you note first
We just crossed 4,850 stars and 1,100 forks on GitHub, with 55 contributors shipping features, fixes, and issues that are actively shaping the roadmap. The Slack community is now 640+ builders strong. This has been humbling - and we're just getting started. Thank you.
→ Star us on GitHub
|
v1.42.0 is out - speech-to-speech gets a big upgrade
We tagged v1.42.0 of the open source stack, and the headline is speech-to-speech: enhancements across Gemini Live, OpenAI Realtime, and Grok, to name a few. One model listens and speaks - no separate STT/TTS hop in between. If you're self-hosting an earlier version, this is the release to update for.
→ Update your self-hosted stack
|
Self-serve credits - buy once, use anywhere
You can now top up credits yourself on Dograh Cloud, no back-and-forth required. Buy on app.dograh.com/billing and spend them in the cloud app - or generate a service key there and drop it into your self-hosted deployment to use Dograh AI models and tokens from your own stack. Same balance, wherever you run.
→ Add credits
|
The MCP server got a lot smarter
The Dograh MCP server can now create and edit your voice agents for you, and build the tools they call. A request like "add a tool that looks up an order and use it in my support agent" becomes a couple of MCP calls instead of manual wiring. Anyone can use it - it works with Dograh Cloud and with your self-hosted instance alike.
→ Connect with the MCP server
|
Share your Dograh story, get up to $200 in credits
Between the MCP breakthrough and the revamped UI, we believe Dograh now offers the most powerful MCPs in the voice AI ecosystem - you can build out just about any voice agent use case in minutes. We want more builders to see that.
Post about your Dograh experience - LinkedIn, X, YouTube, wherever you write - and we'll add a minimum of $20, up to $200 in credits to your account. Reply to this email with the link and we'll also boost it from Dograh's official channels.
|
|
Also shipped this month
More model providers - mostly community-built. Azure AI is now a first-class provider across TTS, STT, embeddings, and realtime. Cartesia lands with Sonic 3.5 for TTS and Ink-2 STT with enhanced turn detection. Inworld, Smallest AI, and Sarvam (LLM plus custom TTS voices) join the roster too - most of these came from external contributors.
Answering Machine Detection. Twilio AMD is now wired into telephony config, so outbound campaigns can tell a human from a voicemail and act accordingly.
Revamped UI. A refreshed look across the app, including a smoother onboarding flow for new workspaces.
|
|
Making it easier to self-host and deploy Dograh
A big theme this month was cutting the friction out of running Dograh yourself:
One-command Docker start. A setup script brings the whole stack up locally, and generates secrets like your Redis password for you.
Free trusted HTTPS for remote installs. Deploying to a public-IP box now gets valid HTTPS out of the box via sslip.io - no domain or cert wrangling to place a real call.
Hostinger deployment files. Managed-Traefik deploy files for a quick, low-cost hosted setup.
Bring your own S3. Custom S3 endpoint, signature version, and addressing style - so MinIO, Cloudflare R2, and other S3-compatible stores just work.
→ Deploy your stack
|
|
How to access these
On app.dograh.com? You're already on the latest - nothing to do.
Self-hosted? Update to v1.42.0 - pull the new images and restart. The update guide covers both "latest" and pinned-tag flows.
→ Update your self-hosted stack
|
|
|
If you try any of this and something feels off - or if there's a model, provider, or integration you want next - hit reply. I read every message.
- Abhishek
Co-founder, Dograh
github.com/dograh-hq/dograh ★
|
|
