AI Voice Agent for Asterisk: Seeking a Frontend Co-Builder

I’ve been building a fully open-source AI Voice Agent for Asterisk/FreePBX and I’m finally at the stage where it needs a proper interface. The backend works great — real-time AI conversations, call handling, extensions, transfers — but the setup is still very CLI-heavy.

I’m comfortable with Asterisk, FreePBX, Linux, and backend systems…
…but not with frontend/UI work.

So I’m looking for a tech-savvy frontend developer who wants to help build:

  • A clean setup wizard

  • A configuration dashboard

  • Real-time call monitoring UI (Currently built with prometheus/grafana)

  • Logs, metrics, and agent management screens

The whole project is open-source (MIT), and while I can’t offer compensation, I can offer:

  • Real, impactful contributions to a fast-moving AI+Telephony project

  • A chance to shape the first fully self-hosted Asterisk AI Voice Agent

  • Visibility and credit on the repo + docs

  • A fun and genuinely useful build for anyone who loves VoIP and AI

If anyone is interested, curious, or just wants to check it out, here’s the repo:

:backhand_index_pointing_right: https://github.com/hkjarral/Asterisk-AI-Voice-Agent

Feel free to DM me — I’d love to collaborate.

The problem with using Cursor for development is that you should set up rules to exclude committing your API keys, passwords, etc to GitHub before you start developing.

If you do accidentally commit secrets, api keys etc upstream, simply deleting the files in the next commit does not clear your commit history.

Finally, you should probably secure your development asterisk server ARI interface with some basic firewalling to control ingress.

Actual project related question: why did you switch from attempting to do both ExternalMedia and AudioSocket to just AudioSocket

On Thu, 20 Nov 2025 at 15:43, hkjarral via Asterisk Community <notifications@asterisk.discoursemail.com> wrote:

hkjarral
November 20

I’ve been building a fully open-source AI Voice Agent for Asterisk/FreePBX and I’m finally at the stage where it needs a proper interface. The backend works great — real-time AI conversations, call handling, extensions, transfers — but the setup is still very CLI-heavy.

I’m comfortable with Asterisk, FreePBX, Linux, and backend systems…
…but not with frontend/UI work.

So I’m looking for a tech-savvy frontend developer who wants to help build:

  • A clean setup wizard

  • A configuration dashboard

  • Real-time call monitoring UI (Currently built with prometheus/grafana)

  • Logs, metrics, and agent management screens

The whole project is open-source (MIT), and while I can’t offer compensation, I can offer:

  • Real, impactful contributions to a fast-moving AI+Telephony project

  • A chance to shape the first fully self-hosted Asterisk AI Voice Agent

  • Visibility and credit on the repo + docs

  • A fun and genuinely useful build for anyone who loves VoIP and AI

If anyone is interested, curious, or just wants to check it out, here’s the repo:

:backhand_index_pointing_right:https://github.com/hkjarral/Asterisk-AI-Voice-Agent

Feel free to DM me — I’d love to collaborate.


Visit Topic or reply to this email to respond.

You are receiving this because you enabled mailing list mode.

To unsubscribe from these emails, click here.

Agreed, that’s a lesson I learned early on to make use of .env and good thing is openai and their are other services that send you an email like within 2-3 minutes that api key has been detected in your github repo, so I did end up deleting/recreating keys few times.
I have good controls in place for my dev environment. Thank you for heads up !
Finally, I started with external media then learned latency lesson and figured audiosocket would work better for full agents like deepgram and gemini live and ended up building both. The project currently works with both external media and audiosocket for full agents or custom pipelines.

Learn to set up one of these in your repository.

I have setup .gitignore in my repo. Thank you for the link :slightly_smiling_face: