Babytalk: Offline speech to text and text to speech on ESP32
Get picks like this daily. The day's top launches, AI/tech news, and a weekly opportunity spotlight — straight to your inbox.
This is 1 of 30 launches in speech processing software — see how it stacks up on momentum and crowding →
188 other launches read as similar to this one →
Details
- External ID
- 50026819
- Source
- HN
- Company
- —
- Product
- Babytalk: Offline speech to text and text to speech on ESP32
- Website domain
- github.com
- Launched
- Oct. 9, 2026
- Cohort
- —
- Upvotes
- 14
- Upvotes percentile
- 0.63
- Tags
- —
Description
I have been playing with a lot of $20 devices with no keyboards or screens. I wanted to use speech to interact with them, but without it having to run through a cloud service or use a fixed command vocabulary.I worked with Claude to implement high performance 4bit int kernels for the ESP32S3 and P4 and quantize some larger STT and TTS models. Surprisingly, it's quite usable on these tiny boards.Also includes a finetuned model that can better cope with my diesel drones, hacking coughs, and mumbling. Works with Micropython (RAM scarcity issues) and AtomVM (better).
Enrichment
- Niche
- speech processing software
- Vertical
- Horizontal
- Function
- Dev tools
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- offline speech to text and text to speech library for esp32 microcontrollers
- Manually corrected
- False
Could you build this?
No Writing custom 4-bit quantized neural network inference kernels in assembly or low-level C for ESP32 hardware requires specialized embedded systems and micro-architecture expertise.
What it would actually take: The project requires custom C/C++ and ESP-IDF assembly to hand-optimize SIMD instructions (ESP32-S3 PIE instructions) for 4-bit integer matrix multiplication. It requires deep knowledge of microcontroller memory hierarchy, DMA transfers, and specialized embedded ML quantization frameworks to fit both STT and TTS models into minimal SRAM.
Discussion
4 comments analyzed.
Concerns raised: unclear use case and purpose, unclear how text output exits device, requires code customization to adapt
Feature requests: hardware STT-to-HID support, clearer README documentation
Competitors
Other products that read as similar to this one — 188 launches clear the similarity bar, closest 8 shown.
Attention rank: #76 of 189 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 338 days after the earliest competitor.
- esp_ai · github · 2026-09-19 · 15 upvotes · similarity 0.50
- Yap · hn · 2026-07-27 · 104 upvotes · similarity 0.47
- ito · github · 2026-10-02 · 11 upvotes · similarity 0.47
- Sidebar · hn · 2026-08-07 · 16 upvotes · similarity 0.46
- TalkiTo · hn · 2025-11-14 · 5 upvotes · similarity 0.45
- F32 · hn · 2025-11-19 · 302 upvotes · similarity 0.44
- Three new Kitten TTS models · hn · 2026-03-19 · 561 upvotes · similarity 0.44
- Text to Speech Local · ph · 2026-10-06 · 1 upvotes · similarity 0.44
Other launches for this product
- No other launches for this product.