This sounds very promising. Using multiple CC instances (or mix of CLI-agents) across tmux panes has always been a workflow of mine, where agents can use the tmux-cli [1] skill/tool to delegate/collaborate with others, or review/debug/validate each others work.
This new orchestration feature makes it much more useful since they share a common task list and the main agent coordinates across them.
I didn’t see that but I do get a lot of stutters (words or syllables repeated 5+ times), not sure if it’s a model problem or post processing issue in the Handy app.
I’m curious about this too. On my M1 Max MacBook I use the Handy app on macOS with Parakeet V3 and I get near instant transcription, accuracy slightly less than slower Whisper models, but that drop is immaterial when talking to CLI coding agents, which is where I find the most use for this.
Since Llama.cpp/llama-server recently added support for the Anthropic messages API, running Claude Code with several recent open-weight local models is now very easy. The messy part is what llama-server flags to use, including chat template etc. I've collected all of that setup info in my claude-code-tools [1] repo, for Qwen3-Coder-next, Qwen3-30B-A3B, Nemotron-3-Nano, GLM-4.7-Flash etc.
Among these, I had lots of trouble getting GLM-4.7-Flash to work (failed tool calls etc), and even when it works, it's at very low tok/s. On the other hand Qwen3 variants perform very well, speed wise. For local sensitive document work, these are excellent; for serious coding not so much.
One caviat missed in most instructions is that you have to set
CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC = 1
in your ~/.claude/settings.json, otherwise CC's telemetry pings cause total network failure because local ports are exhausted.
I absolutely love my UHK [1] split mech keyboard that I ordered from Hungary several years ago. It’s the only one I stuck with after trying some other popular ones.
Other than being split, the keyboard layout is standard so it’s easy to adapt to.
It probably helps me avoid RSI. I keep an apple trackpad between the two splits, so I never use a mouse. And a microphone in the middle as well, you can guess why. I clamshell my MacBook and almost always work on a monitor. Besides ergonomics, the biggest benefit is the on-board programmability; it lets me define custom layers and macros so I can trigger complex window management, app switching, and IDE navigation with simple key combos.
That's exactly how I used to work about 15 years ago, but I found that the Apple trackpad killed my wrists. These days I just have a regular mouse, and simply try to do as much as I can from the keyboard.
I agree trackpad is not RSI-proof by any means, but for me mousing is worse. With the trackpad in the middle I can use either hand to scroll or click etc. I also keep that minimal and instead rely on keyboard tools like Vimium, and scroll kb shortcuts
Thanks , I found it after clicking through to the actual nature paper, where it’s a detail buried deep down in the paper. They really should have mentioned it up front.
I think some of the distinction here is that the more recent "bare LLMs" have been more purpose built, augmented with "agent" specific RL, and in general more fine tuned for the requirements of "agents". Things such as specific reasoning capabilities, tool calling, etc.
These all make the "bare LLMs" better suited to be used within the "agent" harness.
I think the more accurate term would be "agentic LLMs" instead of calling them "agents" outright. As to why its the case now, probably just human laziness and colloquialisms.
GPT 5.2 in a simple while loop runs circles around most things right now. It was released barely a month ago and many developers have been on vacation/hibernating/etc. during this time.
I give it 3-4 more weeks before we start to hear about the death of agentic frameworks. Pointing GPT5+ at a powershell or C#/Python REPL is looking way more capable than wiring up a bunch of domain-specific tools. A code-based REPL is the ultimate tool. You only need one and you can force the model to always call it (100% chance of picking the right tool). The amount of integration work around Process.Start is approximately 10-15 minutes, even if you don't use AI assistance.
My definition of agent has always been an LLM with "effectful" tools, run in a loop where the LLM gets to decide when the task is complete. In other words, an LLM with "agency".
Parakeet V3 is near-instant transcription, and the slight accuracy drop relative to the slower/bigger Whisper models is immaterial when talking to AIs that can “read between the lines”.
This is not strictly speech-to-speech, but I quite like it when working with Claude Code or other CLI Agents:
STT: Handy [1] (open-source), with Parakeet V3 - stunningly fast, near-instant transcription. The slight accuracy drop relative to bigger models is immaterial when you're talking to an AI. I always ask it to restate back to me what it understood, and it gives back a nicely structured version -- this helps confirm understanding as well as likely helps the CLI agent stay on track.
TTS: Pocket-TTS [2], just 100M params, and amazing speech quality (English only).
I made a voice plugin [3] based on this, for Claude Code so it can speak out short updates whenever CC stops. It uses a non-blocking stop hook that calls a headless agent to create the 1/2-sentence summary. Turns out to be surprisingly useful. It's also fun as you can customize the speaking style and mirror your vibe etc.
The voice plugin gives commands to control it:
/voice:speak stop
/voice:speak azelma (change the voice)
/voice:speak <your arbitrary prompt to control the style or other aspects>
Nice, I’ll have to try it out. They should really make a uv-installable CLI tool like pocket-TTS did. People underestimate just how much more immediately usable something becomes when you can simply get something by doing “uv tool install …”
Hi, so I'm looking for an stt that can happen on a server/cron, that will use a small local model (I have 4 vCPU threadripper CPU only and 20G ram on the server) and be able to transcribe from remote audio URLs (preferably, but I know that local models probably don't have this feature so will have to do something like curl the audio down to memory or /tmp and then transcribe and then remove the file etc).
This new orchestration feature makes it much more useful since they share a common task list and the main agent coordinates across them.
[1] https://github.com/pchalasani/claude-code-tools?tab=readme-o...
reply