Run
Solo agent or orchestrator. Terminal work can run in parallel while GUI actions take the mouse one at a time.
- Provider-aware routing
- Pause, stop and resume
- Bounded permissions
Local-first · Open core · macOS
Give AI a desk.
Keep the proof.
Plaux turns a request into observable work across your apps — with permissions, checkpoints, evidence and memory for the next run.
Data stays local
Open-source core
Native for Mac
ORCHESTRATORPublish verified update
01:42
- 01Read project contextKnowledge + instructions✓
- 02Build and signTerminal · deterministic✓
- 03Verify releaseEvidence collection
- 04Publish outcomeWaiting for evidence
⌘Terminalparallel
✦Mac Useobservable
✓Evidenceverified
Drag the sceneLive execution model · illustrative
Orchestration◆Mac Use◆Provider Bridge◆Checkpoints◆Evidence◆Experience Graph◆
◆◆◆◆◆◆
From prompt to proof
Not another chat.
A runtime for digital work.
Models reason. Plaux gives that reasoning a visible, recoverable execution layer across the tools already on your Mac.
See
Every request begins in chat. Every meaningful operation carries time, state and evidence you can inspect.
- Live Desk
- Execution traces
- Verified outcomes
Learn
Successful routes become local experience graphs, so repeated work needs fewer screenshots and model cycles.
- Comparable runs
- Privacy boundary
- Reusable skills
A digital employee, not another tab
Tell it the outcome.
Watch the job happen.
Plaux can work through the apps you already use: searching, checking, typing, sorting and following up. The task stays visible in chat, and you can pause or stop it at any time.
YOUR REQUEST
Build a clean map of relevant Telegram channels
- 01Open Telegram DesktopUse the signed-in app already on this Mac✓
- 02Search topic by topicHuman-paced navigation through the native interface✓
- 03Check relevance and duplicatesKeep only public, useful matches✓
- 04Write verified rows immediatelyThe working document stays up to date✓
RESULTA ready-to-use market map instead of hours of manual searching.verified
Why it can cost less
Use AI for judgment.
Use fast routes for repetition.
A model-only desktop agent repeatedly looks at the screen and reasons about every click. Plaux remembers successful routes and combines them with deterministic tools, so the model is called where judgment is actually needed.
In one tested desktop workflow, the builder observed roughly 100× lower cost than a model-only run. This is a single workload, not a universal guarantee.See→Think→Click→Repeat
Plan→Run→Verify
Tested in real work
Early outcomes, stated honestly.
These are builder- or client-reported experiments, not promises of identical results.
Topic-relevant Telegram entities catalogued
A client used the resulting dataset for competitive targeting and reported an uplift of about ₽3M. Not independently verified.
Market research · client-reportedA social account kept moving
Plaux navigated X, joined relevant conversations and gained 8 organic followers in one unattended experiment.
Social operations · results varyServer and domain taken live
Infrastructure, TLS, domain routing and the product website were configured through one visible task flow.
Launch operations · verifiedA sales document finally became workable
A large document the team struggled to systemize was cleaned, sorted and turned into a usable operating structure.
Back office · verified workflow◎ Plaux works through standard app interfaces where APIs are missing. Operators remain responsible for service rules, permissions and the work they authorize.
Experience Graph
The system gets better at the work — not just better at talking.
Plaux compares compatible runs by reliability, time and verified quality. The graph stores reusable topology and measurements, never your prompt, contacts or file contents.
RELIABILITY92%↑ 18%
AVG. TIME7.9s↓ 3.4s
MODEL LOOPS04↓ 61%
Built in the open
Start local. Grow into a workgraph.
One transparent foundation, with a professional layer being assembled around collaboration, governance and scale.
OPEN SOURCE
CoreCommunity
A local-first macOS workspace for builders who want to inspect, adapt and own their agent runtime.
- ✓ Local projects, chats and files
- ✓ Solo agent and orchestration foundations
- ✓ Cyber Hands with provider permissions
- ✓ Checkpoints, traces and local memory
IN DEVELOPMENT
BuildingPro
For professionals and teams that need shared workflows, deeper controls and measurable business outcomes.
- ＋ Cross-device continuity
- ＋ Team spaces and permission graph
- ＋ Workflow analytics and evidence
- ＋ Curated professional skill packs
IV
Independent developer
Built by Ilia Volkov.
Designed for people who ship.
Plaux started from a practical frustration: capable models still lose time, context and money while operating real software. Ilia is building the missing execution layer — local, inspectable and shaped by real work instead of demos.
Honest roadmap
What exists. What comes next.
Mac Use, projects, orchestration, evidence, provider bridge and local experience graphs.
Deeper workflow branching, comparable evals and three end-to-end agency processes.
Cross-device sync, permissions, team spaces and shared professional workflows.
Plaux Community
AI should finish
the work.
The open core is being prepared for its public source release. Follow the build, test real workflows or help shape the contributor contract.