Stoked y’all are taking a swing at this, tmux isn’t quite right for this, particularly for prioritizing which agents need focus. It’d be helpful to see a walkthrough/screencast of someone’s flow to see some of the typical usage patterns too!
I think Ctrl-w is bound to going back to the agent list, which steals the typical “delete the previous word” shortcut 🫠
13
a guy is quietly billing landscapers $1,247 per ad. each render costs him 96 cents in api credits.
i opened the repo because something in his own screencast contradicted the pitch. at 0:17 claude's own reply said "best model: seedance 2.0." claude isn't making the video. it's running dispatch.
the whole loop:
> drop the client's logo into a wrapper plugged into claude opus
> claude writes a 3 scene shotlist, 5 seconds each, 15 total
> the prompts get fired into veo 3.1 and seedance 2.0 in parallel, claude never touches a frame
> ffmpeg quietly stitches the 3 clips into a 16:9 and a 9:16 mp4
> total spend on the run was 9.6 credits, under a dollar at current pricing
he's not selling a video. he's selling the 30 minutes a small landscaping company would have spent calling a freelancer who would have charged $1,247 anyway.
when the cost to produce a "cinematic" local ad drops to 96 cents, the agency model doesn't die. it just moves to whoever cold emails the 4,847 landscapers in the country first.
the leverage isn't the repo. it's that the people paying $1,247 don't know what veo 3.1 is.
1
59
First semi-proper video screencast of @_checkMono where I show how you add 5 items from a furnished rental just by taking photos and how you inspect them right away to rate their condition for future inspections.
The super power of Mono is reusability and higher frequency through ...
1
49
youtu.be/76YE8ab2Rtk?is=i3vb…
Theory of everything Screencast.
1
38
Some of us are using devices that can actually screencast. E shock you? 🤣🤣🤣🤣
1
414
Jun 13
I was just scrolling the feed, and I see someone recorded a screencast of something.
I look closer - it looks EXACTLY like one small pet project I made for myself (to test some polymarket betting strategies).
I thought - oh shit did he signup and i did not get a notification??
Then i looked closer and figured it out. AI just builds absolutely same UI for everyone. Layouts. Colors. Fonts. Everything is the same.
1
23
Jun 13
That's WILD. I don't have any of my old n64 stuff. I use an emulator on my phone and then screencast it to my TV to play any old games!
1
1
18
Jun 12
maxstream tv ajg kg bisa di screencast gmn ini ajrrr udh bayar masalahnya 🤬🤬🤬
71
Jun 11
Guy how I go fit reach you nau jhoorrrrr.. I have the screenshot and screencast
2
208
Jun 11
You stop screencast just before «we kidnapped your kids» paywall
6
1,246
Jun 11
can you DM me details of what happened? screencast or any more details would help.
happy to figure out what's going on!
7
2,775
Jun 10
Tune Lab Q 1.0.1 Hotfix 🔥
Download "Tune Lab Q 1.0.1"
GD: drive.google.com/file/d/1Vzz…
TG: t.me/tune_lab_q/7
This hot fix focuses on stability, crash diagnostics, safer extension loading, and a small auto-scroll behavior improvement.
🛑What changed
— Added improved crash logging.
— Added better error logging for export/render issues.
— Added safer handling for missing audio files in projects.
— Improved stability when editing notes, splitting notes, editing lyrics, playing back, and rendering/exporting.
— Improved internal handling of synthesis pieces to reduce possible random crashes during editing/playback/rendering.
— Improved extension and voice engine loading:
— built-in engines are now loaded before user extensions;
— duplicate engine types are skipped;
— extension loading errors are logged more clearly.
— Disabled original TuneLab update checks inside TLQ.
— TLQ now uses its own app data folder instead of sharing the original TuneLab folder.
— Auto-scroll is now enabled by default for new users.
— Added more detailed logs around export/render operations.
— Important note about folders
— TLQ now stores its data separately from original TuneLab.
🛑Running TLQ as administrator may reduce crashes for some users. But it is not the recommended permanent solution.
🛑New TLQ folder:
%APPDATA%\TuneLabQ
Logs folder: %APPDATA%\TuneLabQ\Logs
This change is intentional. It helps prevent conflicts with original TuneLab settings, extensions, bridges, and cached data. Because of this, some settings or extensions may need to be set up again after updating.
🛑If TLQ crashes
Please send a crash report with the following information.
1. Log file
— Open TLQ and go to: More → Open Tune Lab Q Folder
— Then open: Logs
— Send the latest log file. It should look similar to: TuneLabQ_2026-06-10_19-00-00.log
You can also open the folder manually by pasting this into Windows Explorer:
%APPDATA%\TuneLabQ\Logs
2. What you were doing before the crash
Please describe the last action before TLQ crashed, for example:
— splitting notes;
— editing lyrics;
— changing phonemes;
— playback;
— render/export;
— importing SVP/VPR/VSQX/VSHp;
— changing singer/voice;
— using XSY;
— opening a specific project.
3. Project file
If possible, send the project file where the crash happened.
This is especially important if the crash happens repeatedly with the same project.
4. Voice / bridge information
Please tell me:
— which singer/voice was selected;
— which bridge/engine was used;
— whether the crash happens with one specific voice or with different voices.
5. System information
Please include:
— TLQ version:
— Windows version:
— CPU:
— RAM:
— Audio device/interface, if relevant:
6. Screencast, if the crash is repeatable
If you can reproduce the issue, a short screencast is very helpful.
Please try to show:
— the action right before the crash;
— the selected voice/track;
— the project area where it happens;
— any error message, if one appears.
🛑Please make sure TLQ is extracted into a normal user-writable folder, for example:
C:\Users\<YourName>\Apps\Tune Lab Q\
Avoid running TLQ directly from protected folders such as: "C:\Program Files\"
Also avoid running it directly from inside the ZIP archive.
Jun 10
Oh hi🥺
I just released Tune Lab Q
It’s my reimagined version of the original TuneLab 1.5.4, made from the perspective of a long-time vocal synth user.
What started as a small personal UI mod eventually grew into something much bigger. The goal of TLQ now is to make tuning (especially VOCALOID tuning) more comfortable, flexible, and visually enjoyable.
TLQ comes with a fully redesigned dark interface, a new workflow, improved pitch and dynamics editing, auto-pitch presets, interface customization, global XSY morphing, custom sample loading, SV2 .svp import with pitch phoneme timing, VocalShifter project support, and a lot of small improvements that make tuning feel better. You can even cross-synthesize voices between different engines - for example, Miku V6 with Luka V2.
Download links on YT: youtu.be/CNLvlizcRr0?si=uhZa…
I would be really grateful if you try it, share your works, and send feedback, bugs, or suggestions!
Cover in the showcase
BE NO BODY from @alissicofficial [JPN version]
Singer: 初音ミク V6 x V4x (with TLQ morph-system)
#VOCALOID #TuneLabQ #TLQ #初音ミク
3
12
71
6,227
Jun 9
@AwesomeMotive @syedbalkhi I have a bug to report for EDD/EDD Software Licensing add-on. Where can I send a screencast showing the issue?
1
53
Jun 9
2. AI Reframe 2: layout magic
Klap’s new update analyzes your scenes and applies perfect layouts:
→ Split screen for interviews
→ Screencast for tutorials
→ Zoom for gaming or podcasts
Each clip looks like it was made by a pro.
1
374
Jun 8
day 8/30
90% of the AI agents being built right now are completely useless in production.
it’s incredibly easy to make a multi-agent system look like magic in a 1-minute screencast on X. but the moment you hand it to real users, it falls apart.
after 2 years of deploying models to production at a singapore firm, i’ve realized successful orchestration comes down to three things most devs ignore:
:) killing the LLM loop early: if a rule can be hardcoded with a simple if/else script, do not waste tokens letting an agent guess.
:) aggressive context pruning: passing an entire conversational history back and forth is just distributed prompt bloat.
:) semantic caching: caching common tool-calling payloads saves massive latency and cloud budget.
stop building fancy prototypes that look cool but cost a fortune to run. focus on deterministic infrastructure that actually scales.
what’s the biggest bottleneck you’ve run into trying to take an agent to production? let's talk 👇
#LLMOps #AIAgents #GenAI #BuildInPublic
2
1
2
38