LocalLM Studio and private runtimes
IMPLEMENTEDLocal-first AI operations
Route the work.
Not just the prompt.
LM Manager Pro turns models, machines, personas, teams, chat, and image engines into one coherent creation studio—without giving up control of where the work runs.
GitHub access is required to view this private repository.
CreateLibraryActivitySettings
CREATE
What do you want to make?
Describe the outcome first. Review one ready route before work begins.
Draft a private product concept and compare the strongest result.
CHATLocal or cloud
IMAGEApple or selected engines
TEAMReview configuration
ROUTE OPTIONS
On-device · Local · Cloud
readiness checked before launch
On-deviceApple-native model paths
IMPLEMENTEDCloudProvider routes when chosen
OPTIONALVersion 2.28.0 · Released August 30, 2026
Private creation is live. The next direction stays evidence-led.
LM Manager Pro is live on the App Store for iPhone and iPad. The shipped app centers four destinations—Create, Library, Activity, and Settings—and gives Chat, Image, Compare, Team, and Episode work a clear starting mode.
-
Released
Version 2.28.0
Create leads with clear Chat, Image, Compare, Team, and Episode modes while keeping the selected intent visible through route review.
-
Now
One visible home for each capability
Create starts the work, Library owns saved creations and reusable assets, Activity follows current runs and evidence, and Settings owns configuration.
-
Next
Keep every route truthful
Continue hardening live-model readiness, setup repair, tool-assisted runs, and physical-device acceptance without hiding incomplete configuration.
-
Later
Prove before expanding
Close real-engine and accessibility acceptance, preserve migration coverage, and add no hidden context, silent route changes, or fictional execution state.
Released, with evidence boundaries intact.
The live App Store listing proves public availability, not every source-provenance or hardware claim. Repository evidence records focused tests, builds, release assets, and the remaining real-engine and historical tag limitations explicitly.
From intent to measured result
One operating loop for every kind of AI work.
Describe the outcome, review a ready route, launch it across the right resources, then keep the people, assets, history, and performance signals together.
Start with the work
Use chat, a reusable persona, a team, an episode, a comparison, or an image workflow without rebuilding context each time.
See what is ready
Create separates route readiness, machine state, workflow compatibility, and provider capability before handing work to Chat, Image Studio, Compare, or a coordinated workflow.
Measure what happened
Keep history, reusable assets, latency, throughput, quality, costs, and advisor signals attached to real runs.
Creation without a provider maze
Chat, images, video routes, teams, and reusable runs.
- Machines & modelsTrack local LM Studio hosts and model availability.
- Image StudioRoute across Apple, local diffusion engines, ComfyUI, and selected cloud providers.
- Personas & teamsTurn collaborators and structured conversations into reusable building blocks.
- ActivityFollow current work and compare measured results instead of relying on model names alone.
The operating principle
Your hardware stays useful. Your choices stay visible.
LM Manager Pro is local-first, but not local-only. It makes the execution route explicit so privacy, capability, speed, and cost remain decisions—not hidden defaults.
Compare the experiments