Muslim Platform Deploys Arabic Voice AI for Sourced Islamic Knowledge
Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge
The value here is less the models than the operational playbook: abuse-resistant metering, deferred email verification, and observability designed for a GPU host failing silently. Engineers running voice agents in production can copy this stack directly. Arabic TTS performance on the open TTS Arena also gives a concrete benchmark against alternatives.
Muslim is a production Arabic voice AI platform delivering sourced Islamic knowledge through a real-time voice pipeline built on NeMo Arabic ASR, an OpenAI-compatible LLM endpoint, and self-hosted TTS.
The team released fine-tuned model artifacts, including Muslim-6B-PRO, a 5.94B-parameter tool-routing LLM, and Fasih-TTS-V1, which ranks 5th of 17 overall and 2nd of 11 open-weight systems on the community-voted Arabic TTS Arena for Modern Standard Arabic.
Retrieval runs across six MCP servers in a deterministic multi-source routing layer.Measured figures include 98.
4% recitation-validation accuracy on 124 cases and end-to-end voice latency of 0.9-1.
7 seconds.The paper also details an account and metering layer plus a three-layer observability stack for the system's main failure mode: a GPU-bound agent host going silent while the web tier keeps serving.
Muslim: A Deployed Arabic Voice AI Platform for Grounded Islamic Knowledge
We present Muslim, a production Arabic voice AI platform serving grounded, sourced Islamic knowledge to real users. Beyond a real-time voice pipeline (NeMo Arabic ASR, an OpenAI-compatible LLM endpoint, self-hosted TTS) and a deterministic multi-source retrieval layer routed across six Model Context Protocol servers, we report three things a research prototype typically lacks. First, a released family of fine-tuned Arabic Islamic model artifacts: an efficient tool-routing LLM (Muslim-6B-PRO, 5.94B parameters) and a Modern Standard Arabic TTS model (Fasih-TTS-V1) that ranks 5th of 17 overall and 2nd of 11 open-weight systems on the community-voted Arabic TTS Arena for MSA. Second, an account and metering layer - a free per-account turn allowance, capacity-aware refusal, and email verification deferred to the point it actually matters - that turns an open demo into an operable, abuse-resistant product. Third, a three-layer observability stack (liveness, error reporting, product analytics) built specifically around the system's characteristic failure mode: a GPU-bound agent host going silent while the web tier keeps serving normally. We report real, measured latency and accuracy figures (98.4% recitation-validation accuracy on 124 cases; end-to-end voice latency of 0.9-1.7s) and discuss the concrete engineering trade-offs and limitations of running an Islamic-knowledge voice product in production.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Muslim, MCP and tools, speech recognition, TTS.
Primary source verification: published directly by arXiv: OpenAI with first-party authority. Cross-source corroboration: 11 independent sources have covered this development (including lmarena.ai, kimmonismus, OpenAI), confirming industry-wide signal strength.
