The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
This is agrntfuse dashboard I built for detecting whent the agent goes rogue, drifted from actual goal or looping or doing something random work not actual task, so it dtects and tell how muc it cost the tokens and how much amount has been used up, and steer it back to the ctaula goal with a better reasoning model.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
CRAG — Self-Correcting Research Assistant
Overview
A research-assistant project exploring how AI applications can retrieve information and improve the quality of their responses through a self-correction workflow.
Project focus
Python-based AI application development.
Retrieval and research workflow design.
Improving answer reliability through evaluation and correction.
Technologies: Python, AI, retrieval-augmented generation.
Self-correcting retrieval is the upgrade RAG has needed, Daniel. Most systems stop at the first retrieval and hope for the best, so building evaluation and correction into the workflow is a real reliability win. The live telemetry dashboard makes it feel production-ready.
Sub-400ms voice agent: Eliminating audio lag in conversational AI 🎙️⚡
Most voice AI bots feel robotic not because of the LLM's intelligence, but because of the latency penalty from traditional HTTP chaining: Audio In ➔ Whisper API ➔ LLM API ➔ ElevenLabs API ➔ Audio Out. That round-trip easily takes 2.5 to 4 seconds, killing natural conversational flow.
Currently building an ultra-low-latency real-time voice pipeline using direct audio streaming:
Bidirectional WebRTC Data Channels: Replaced standard REST request-response cycles with duplex WebRTC streams, piping raw PCM audio chunks continuously.
Streaming VAD (Voice Activity Detection): Implemented Silero VAD at the edge to catch user interruptions instantly, immediately sending cancellation tokens to abort in-flight LLM token generation.
Token-to-Speech Streaming: Connected streaming LLM token generation directly to a streaming TTS synthesizer, firing synthetic phonemes while the sentence is still being composed.
Context Preservation: Managing conversational state via low-latency Redis cache so context remains intact even across frequent back-and-forth interruptions.
Achieving natural, fluid interruptions and dropping perceived latency down to human conversational speeds (~380ms).
Architecture flow and latency breakdown below! 👇
#VoiceAI #WebRTC #Python #FastAPI #AudioEngineering #GenerativeAI #BuildInPublic
Market Day — the 15-second ad every seller deserves
Small sellers pour everything into what they make, and then get two seconds of someone's scroll to show it.
Market Day is for that moment. A seller ties off a parcel, the stall lights flicker on, and they hold out a wooden tray toward you. Your product lands on it, a little kraft label hops up to say what it is, and your headline settles onto a frosted-glass card.
I built it as one SVGator template and dropped in three businesses to prove it bends without breaking:
🍕 Forno Lucia, a wood-fired pizza café
🕯️ Ember & Oak, a candle maker
☕ Clay & Co, a ceramics studio
Same animation every time. To make it yours, swap 5 lines of text, 2 images and 1 colour. Every layer is named, the motion lives on wrapper groups, and no keyframe needs touching. About five minutes.
All artwork was made for this template, the brands are fictional and the fonts are open source, so it's free to give away.
If you sell by hand, this one's for you. Which stall should I set up next? 👇
Template: https://app.svgator.com/editor/#/0c516a7c8d234b37bb98c8df7f0b4a29
@SVGator
