1. Conduid
  2. Communication
  3. Saybench
MCP server · Communication

Saybench

by · github.com/renan-martini/saybench

Benchmark the full voice-AI pipeline on your own audio — STT (batch + streaming), LLM TTFT, speech-to-speech, and TTS. WER by failure mode, keyterm recall, real-time-paced latency, cost columns, a CI regression gate, and an MCP server for coding agents.

Unclaimed benchmarkingclideepgramgogolanglatencyllmmcpmcp-serveropenaipipecatspeech-recognitionspeech-to-speechspeech-to-textstttext-to-speechttsvoice-agentsvoice-aiwer
InstallView source
34Low

Scored a month ago · breakdown

About Saybench

Saybench is an MCP server published by renan-martini in the Communication category: benchmark the full voice-AI pipeline on your own audio — STT (batch + streaming), LLM TTFT, speech-to-speech, and TTS. WER by failure mode, keyterm recall, real-time-paced latency, cost columns, a CI regression gate, and an MCP server for coding agents.

The repository has 1 stars and 0 forks, with the last commit a month ago. Recent commit activity is the strongest signal that reported issues will get attention.

Install

From source
git clone https://github.com/renan-martini/saybench

This server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.

Ask AI

✦

Ask AI about Saybench

Powered by Claude · Grounded in docs

I know everything about Saybench. Ask me about installation, configuration, usage, or troubleshooting.

Security checks

6 checks have not been run yet.

Releases

v1.1.0v1.1.0 — the deferred items, un-deferred · 9 Sep 2026The three items v1.0 deferred with a rationale, now shipped: `-judge <llm-target>` (on `stt`, `stream`, and `s2s -score echo`): an LLM rates each transcript against its reference on a fixed rubric — does the meaning survive? — and a…
v1.0.0saybench 1.0 · 9 Sep 2026Benchmark your voice-AI stack on your own audio, prompts, and providers — six surfaces, one honesty regime. Batch + streaming STT: WER by failure-mode category, keyterm recall, TTFP/finalization lag, interim word survival LLM…
Questions

About Saybench

How do I install Saybench?

No package has been published for this server, so install it from source: clone the repository and follow its README.

Is Saybench safe to use with an AI agent?

Its trust score is 34 out of 100 (low). Conduid hasn't run static security checks on this repository yet, so review the source yourself before granting it credentials. It has no ConduID identity yet, so agent calls to it are not receipted.

Is Saybench still maintained?

Yes — the latest release is v1.1.0 (9 Sep 2026), and the last commit was a month ago. The repository has 1 stars.