sunona voice agent.

Next-generation voice AI infrastructure. Orchestrate intelligent, action-oriented voice agents that sound human and execute complex workflows via function calling.

Comment

Sunona AI

Inspiration

Sunona AI was inspired by the vision of creating an open‑source voice intelligence platform that blends backend stability with real‑world usability. We wanted to democratize access to advanced voice AI without locking it behind commercial SaaS barriers.

What We Learned

  • How to design scalable backend infrastructure using PostgreSQL, Redis, and Docker.
  • The importance of clear documentation and reviewer‑friendly ASCII diagrams.
  • How licensing can protect open‑source projects from misuse.

How We Built It

  1. Backend: Developed with Python + FastAPI, containerized via Docker.
  2. Database: PostgreSQL hosted on free tiers (Neon/Supabase).
  3. Cache Layer: Redis for fast response times.
  4. Deployment: Optimized on Railway/Render free tiers with Cloudflare SSL.

Challenges

  • Limited budget for hosting production workloads.
  • Worker crashes (Uvicorn restarts, Redis startup issues).
  • Tight page limits for hackathon submissions, requiring concise ASCII diagrams.

Example Markdown Formatting

  • Bold text for emphasis
  • Italicized text for softer highlights
  • > Blockquotes for key insights
  • Lists (ordered/unordered) to keep content structured
  • Inline code for commands or snippets

Example LaTeX Math

Inline (Latency):
Average response latency L = (t1 + t2 + t3 + ... + tn) / n

Display (Throughput):
T = R / (C * U)

Where:

T = throughput (requests per second)

R = total requests handled

C = number of CPU cores

U = utilization factor

Architecture (ASCII Diagram)

                     +-------------------+
                     |   Client Layer          |
                     | (Web / Mobile UI)  |
                     +-------------------+
                                |
                                v
                     +-------------------+
                     |   API Gateway         |
                     | (FastAPI / NGINX) |
                     +-------------------+
                                    |
    -------------------------------------------------
    |                                                             |
    v                                                            v

+-------------------+ +-------------------+ | Voice Processing | | Auth & Security | | (Speech-to-Text, | | (JWT, OAuth) | | NLP Modules) | | | +-------------------+ +-------------------+ | | -------------------+----------------------------- | v +-------------------+ | Core Backend | | (Sunona Engine) | | Business Logic | +-------------------+ | ----------------------------------------------- | | v v +-------------------+ +-------------------+ | PostgreSQL DB | | Redis Cache | | (User Data, AI | | (Sessions, | | Models, Logs) | | Fast Lookups) | +-------------------+ +-------------------+ | | -------------------+--------------------------- | v +-------------------+ | External Services | | (Cloudflare CDN, | | Storage, APIs) | +-------------------+

Final Notes

We believe Sunona can become a global benchmark for open‑source voice AI, empowering developers, educators, and communities worldwide.

+ 3 more
Share this project:

Updates