Cekura - The self-improvement loop for voice agents

Cekura is the testing, observability, and self-improvement platform for production voice and chat AI agents. It simulates thousands of scenarios, catches failures, diagnoses the root cause, rewrites prompts and config, then re-validates with a full regression sweep. Unlike tools that hand failures back to your team, Cekura closes the loop by fixing the agent itself and proving the fix holds without overfitting.

Add a comment

Replies

Best

Congrats on the launch. The hard part of a self-improvement loop is the blast radius. When an agent rewrites its own behavior from production calls, a fix for one flow can quietly bend a compliance flow sitting next to it, and in healthcare that is the thing every security review hunts for. Closing that loop safely, so the agent improves without drifting out of its guardrails, is the whole game, and it is a genuinely hard problem to have taken on.

 Thank you, really appreciate it

Lets Go Team!!!!

Simulating messy real-world conversations with interruptions, pauses, and background noise feels much closer to production than traditional scripted evaluations.

 Thanks, appreciate it

Love seeing the focus on proving that a fix doesn't introduce regressions elsewhere. Reliable AI systems need repeatable validation, not just faster debugging.

Love the idea of making voice agents “self-healing” instead of just observable.

 thanks! appreciate it.

this is squarely the problem we deal with running voice AI. the thread so far is all about logical correctness (did the agent do the right thing), but a huge chunk of our real failures are cases where the words are technically right and the delivery is off - wrong pacing after an interruption, a flat tone on something that should sound apologetic, talking over a caller who paused to think rather than finished. does Cekura's simulation/scoring catch prosody and delivery quality as a distinct failure category, or is it mainly evaluating on transcript content right now

 Good question. We provide several out-of-the-box metrics for exactly this , pacing, interruption timing, and voice tone/clarity , computed straight from the audio. We also support custom metric creation, including audio-based ones, so you can define your own metric to judge delivery, tone, or pacing beyond what's built in

 good to know pacing and interruption timing are already built-in rather than something we'd have to define ourselves. the case I'm most interested in is the interruption-timing one specifically - talking over a caller who paused to think vs one who actually finished is a judgment call even for a human listening back, since the only signal is a gap length that could go either way. is that metric scored against a fixed silence threshold, or does it try to account for what's actually being said before the pause (trailing off mid-sentence vs a completed thought)

 Hi Gal, regarding your use case for interruption timing during caller pauses: while we don't have a pre-defined metric for this, you can easily build a custom one. Our Metric Builder lets you describe what you need in plain text, and it generates the metric using your recent call data. Since custom metrics analyze both audio and transcripts, your scenario is easily supported. Also we have Metric optimiser feature too where you can use false positives calls to optimise the metric further. Hope this helps.

 good to know it's buildable, but that puts the burden on me to define "completed thought" from scratch in plain text. since you already run transcript analysis alongside audio, does that layer surface anything like incomplete-utterance signals (trailing conjunctions, unfinished clauses) that I could point the Metric Builder at, or would I genuinely be starting from a blank page describing linguistic completeness myself

This is great and a much needed product. How are you processing the simulated calls? Is it STT or do the models listen to the audio stream to pick tone, frustration signals etc.

 Both , we use STT to build the transcript for content/accuracy grading, but for tone, pacing, clarity, and similar delivery signals, models analyze the actual audio stream directly rather than relying on the transcript.

Nice launch! What CI/CD tool integrations do you support?

 Hi Julien, GitHub Actions is the one we support out of the box. For some other CI system there's no separate integration needed: our CLI does the same job. You can check more on CI/CD here:

When monitoring production calls, how does Cekura detect quality issues in real time, and what kind of alerting or reporting does it provide to teams?

 For production monitoring, Cekura runs metrics continuously on live calls and clusters failures into recurring failure-mode themes automatically, so you see patterns instead of re-reading every call. On top of that, configurable alerts (failure, trend drift, threshold breach, new failure mode, failure-mode spike) route straight to Slack with the call and context attached.

Does Cekura support testing across multiple languages and accents, given how critical that is for conversational AI reliability?

 Yes , language and accent are both first-class testing surfaces. Personalities let you run the same evaluator across different languages (including code-switching, like Spanglish) and different regional/non-native accents, with a Transcription Accuracy metric that flags exactly where STT struggles per variant, so you can compare pass rates side by side instead of relying on one aggregate number.