We’ve made meaningful improvements to the performance of the Rime dashboard by simplifying how pages load and how authentication is handled behind the scenes.
What changed under the hood
The core change was centralizing our authentication logic and caching it client-side. By reusing session information and eliminating redundant network requests, the dashboard no longer needs to re-fetch the same data every time you move between pages.
Before this update, each page navigation triggered its own authentication check and a fresh round of data fetching. Those requests added up, and the delay was most noticeable for teams that live in the dashboard while tuning a voice agent. Consolidating that logic into a single cached session removed entire classes of duplicate work.
As a result, navigating between dashboard pages now takes under 50ms. Effectively instant!
Why dashboard speed matters
The dashboard is where most Rime work happens outside the API itself. It’s where you preview and compare voices, create and manage API keys, check word coverage, generate custom pronunciations, and track usage and billing across your team. Building a production voice agent means bouncing between these pages constantly: audition a voice, copy a key, run a test, check usage, repeat.
When each page load takes a second or two, that loop slows down and so does your iteration speed. When navigation is instant, the dashboard stays out of your way.
The gain is biggest for the workflows that touch the most pages. Auditioning voices side by side, rotating an API key before a deploy, or reconciling usage across a team account all involve a handful of quick visits to different pages. At a second per load, that friction was noticeable. At 50ms, it disappears.
Performance is the whole point
Rime builds speech models for real-time conversation, where milliseconds of latency decide whether an AI agent sounds responsive or awkward. We hold our own tools to a similar standard. A team that chooses Rime for low-latency text-to-speech shouldn’t have to wait on a slow dashboard to manage it.
This update follows the same engineering approach we apply to our TTS APIs: measure where the time actually goes, remove the redundant work, and verify the improvement with real production traffic.
The same numbers anchor our model work. Coda, our newest speech model, runs at sub-100ms model latency in production. Mist v3 delivers first audio in about 37ms at the median. Holding the dashboard to a 50ms navigation budget is the same discipline applied to a different surface.
Prefer a terminal? There’s a CLI and an MCP server too
The dashboard is one of three surfaces for working with Rime, and the other two are built for speed in their own ways. The Rime CLI generates and plays audio, inspects usage, and checks latency without leaving your shell, which makes it handy for scripted smoke tests and quick latency spot-checks. The Rime MCP server at mcp.rime.ai connects Claude, Codex, and other MCP-compatible tools directly to Rime, so your coding assistant can manage voices and scaffold Pipecat or LiveKit integration code while you build.
The API key you create in the dashboard works immediately across the CLI, the MCP server, and the HTTP and WebSocket APIs, so there’s nothing extra to configure.
No changes required on your end
No new features to learn. Just a faster, smoother Rime dashboard.
Everything works exactly as before. Your API keys, saved pronunciations, usage history, and team settings are untouched, and the speedup applies automatically the next time you log in at app.rime.ai.
If you notice anything that still feels slow, tell us at help@rime.ai. Performance work is never really finished, and reports from real workflows are the fastest way for us to find the next bottleneck.
More from the lab
Writing for the ear: Prompting your TTS to sound human
Rime was founded by linguists. We spent years studying how people actually talk before we trained models on it, so our prompting guide teaches something different: how to transcribe speech the way it’s spoken. Below is the short version of that guide and gives some context behind why we do it this way.
.png)
Text to Speech for Business: How to Choose a TTS App That Scales
This guide shows you how to pick a TTS app that holds up in production.
What Is Back-Channeling in Conversational AI? A Plain-English Guide
This guide explains back-channeling in conversational AI in plain English. You'll learn what it is and how a voice agent decides when to use it. You'll also see how it produces a natural cue and why timing makes or breaks the effect.