Mesurii — Your AI Lab Partner Inside VS Code
Context-aware, instrument-fluent, and built for LAMB devices. Mesurii reads your repos, terminals, cameras, and notebooks to propose grounded changes — with agentic powers to spawn shells, capture feeds, and edit files safely.
- Code, diffs, and notebooks (File/Notebook panel)
- Live terminals with spawn/attach/resize/kill
- Camera captures from LAMB or your device
- Chat history with folders and revisions
An AI pair-engineer for test & measurement workflows
Mesurii is an operator-first chat surface that sits inside the LAMB landing experience. It streams reasoning, renders tool calls as rich cards, and gives you a side panel for terminals, diffs, files/notebooks, cameras, and image control — all without losing conversational flow.
Operator-first
Quick pivots between chat, terminals, cameras, and diffs, with auto-open listeners that bring the right panel to focus when tools fire.
Agent-forward
Agents can spawn shells, capture images, read files, and apply patches. Tool traces stay inspectable to keep you in the loop.
Photonic clarity
Glassmorphic layers, neon accents, and purposeful typography tuned for legibility across dark lab environments.
Stop fighting instruments. Start running experiments.
- Fragmented context across PDFs, notebooks, half-finished scripts, and terminal history.
- Brittle automation that breaks on DUT or setup changes.
- Generic AI tools that don’t speak oscilloscopes, PSUs, switch matrices, or SCPI constraints.
Mesurii brings everything into one loop
Reads your tests, lab modules, and notes; routes tools through terminals, file APIs, and cameras; proposes grounded edits consistent with your stack. You stay in control by reviewing diffs, patches, and tool traces.
Built for measurement automation, not generic CRUD
Context-aware code assistance
- Understands project structure, virtual envs, fixtures.
- Explains unfamiliar code paths and refactors legacy scripts.
- Generates tests and measurement routines consistent with your repo.
Instrument-aware guidance
- Composes SCPI and instrument APIs with guardrails.
- Suggests sequences for scopes, PSUs, DMMs, AWGs, matrices.
- Turns notebooks and notes into executable procedures.
Agentic surface
- Terminal hub: spawn/attach/resize/kill with live buffers.
- Diffs: repo status, staged/unstaged, commits, undo.
- Files/Notebooks: proxy-aware tree, inline edits, Jupyter kernels.
- Cameras: LAMB server capture + local getUserMedia, attach to chat.
- Images: auto-compress, recompress-on-error, multimodal sends.
Cost-aware, router-backed AI
- Pick between local open-weight models or premium APIs.
- Policies for “fast & cheap” vs “high-accuracy” runs.
- Clear boundaries for secrets and offline-only flows.
From install to running experiments
Connect VS Code to your router (LABIIUM, local model host, or cloud endpoint).
Open your measurement repo, configure envs, and point Mesurii at LAMB devices.
“Write a PyTestLab sweep,” “Debug this SCPI timeout,” “Refactor these notebooks.”
Inspect diffs, tool traces, and terminal output before committing or executing.
Use cases
Lab automation engineers
Standardize test scripts, reduce boilerplate, and onboard teammates faster.
Academic + teaching labs
Guided assistance with governance overlays; hint-only or restricted modes for students.
R&D characterization teams
Turn one-off scripts into reusable pipelines across DUTs and projects.
Instrument & hardware startups
Co-design SDKs and examples with Mesurii inside your repos for better docs.
You stay in charge — of models, budgets, and behavior
- Define which models are allowed per project or cohort.
- Set cost ceilings and prefer local models when budgets tighten.
- Overlay behaviors for “productive struggle” or deeper technical guidance.
Agent access toggles (`agentEnabledCams`) gate autonomous camera calls. Tool events auto-open side tabs (terminal/diff/files) so you can watch agent output without hunting for it.
Captured camera thumbnails can auto-attach to the composer for immediate multimodal prompts.
Questions you might have
Does Mesurii require cloud access?
No. Point Mesurii at any router endpoint: fully local open-weight models, cloud providers, or a hybrid setup.
What languages and frameworks are supported?
Optimized for Python and Rust in test-and-measurement workflows, with first-class PyTestLab and LABIIUM support.
Can I use Mesurii in a teaching lab?
Yes. Configure routing and overlays for students — hint-only modes, restricted models, and teacher dashboards via the LABIIUM router.
How are code and data handled?
Only the context needed for a request is sent to your configured backend. With local routing, code and data never leave your environment.
Do I need LAMB devices or PyTestLab?
No, but pairing with LAMB devices unlocks camera capture, terminal orchestration, and richer automation directly from chat.