{"resourceId":"lbnl-molecular-foundry-tem-agent-mcp-microscope-20261029","versions":[{"version":"external-1a8840ef1399d7586d36b7e0d55ff732f6806ec9fbde797e5d632d6381f9de96","resource":{"id":"lbnl-molecular-foundry-tem-agent-mcp-microscope-20261029","title":"Berkeley Lab Molecular Foundry connects a commercial LLM to an advanced electron microscope through MCP, letting users run experiments in plain language","organization":"Molecular Foundry, Lawrence Berkeley National Laboratory (DOE Office of Science user facility)","sector":"National laboratory user facility and research instrumentation","geography":"United States (California); applicable to university core facilities","publishedAt":"29 October 2026 (Molecular Foundry news, adapted from a Berkeley Lab press release)","publicationDate":"2026-10-29","eventDate":null,"sourceName":"Molecular Foundry, Berkeley Lab","sourceLabel":"Molecular Foundry, \"New Virtual Assistant Streamlines One of the Most Advanced Microscopes,\" 29 Oct 2026","sourceUrl":"https://foundry.lbl.gov/2026/10/29/new-virtual-assistant-streamlines-one-of-the-most-advanced-microscopes/","evidenceClass":"vendor-claim","outcomeClass":"emerging","topics":["developers-agents","infrastructure","operating-model"],"finding":"A Molecular Foundry team rewrote automation routines it had already built for a transmission electron microscope in Model Context Protocol (MCP) format, which gives a commercial language model a fixed set of pre-written actions rather than open-ended control. The resulting \"TEM Agent\" lets researchers describe an experiment in plain language and strings together steps such as changing magnification, tilting the sample and refocusing the beam. In its first tomography attempt it confused sample tilt with scan rotation, which the team fixed with a short written explanation. The team also tested ptychography. TEM Agent is linked to NERSC and to Crucible, the Foundry metadata platform, which the article says lets raw data be processed in seconds instead of hours and lets users ask what settings they used weeks earlier. A single detector can generate about 700 GB in 15 seconds, and visiting users can need weeks of training. The work is described in an npj Computational Materials study.","sledRelevance":"Shared instrument facilities at research universities face the same training and data bottleneck. This shows a lower-risk pattern: expose existing, tested automation to an LLM through a bounded tool interface and a metadata platform, rather than giving the model direct control.","evidence":"Read the full article. No accuracy, error-rate, safety-incident or user-outcome measures are given; the seconds-versus-hours figure is the lab's description. The cited journal paper was not inspected.","architectureImplications":"Bounded MCP tools wrap existing instrument scripts; links to an HPC center and a metadata catalog provide processing and recall of past settings.","governanceImplications":"Restricting the model to pre-written actions is a control choice; facilities should still define which actions need human confirmation.","securityPrivacyImplications":"A commercial model is connected to instrument control and experiment metadata; the article does not describe data handling, authentication or model hosting.","caveats":"Institutional news adapted from a press release; the agent's reliability is described through anecdotes, including one initial failure. Results come from one facility with mature existing automation, which the team says made integration easier than expected.","streamIds":["research"],"roles":{"sales":"Lead with access to shared instruments rather than autonomy. Berkeley Lab's Molecular Foundry reported on 29 October that its TEM Agent lets researchers run an advanced electron microscope in plain language by calling pre-written automation through MCP, linked to NERSC and a metadata platform. University core facility directors with long training queues face the same problem. Offer a bounded pilot on one instrument with existing scripts. Do not promise reliability or training savings; no measured outcomes were reported.","engineering":"Frame fit around wrapping existing instrument automation as bounded MCP tools. The Foundry did not give the model open-ended control; it exposed routines it had already built, added a metadata catalog and connected HPC processing. Prerequisites are scriptable instrument interfaces, an experiment metadata store and an approved model service. Validate on established techniques first, as the team did with tomography, and log every tool call, since its first attempt confused two actions.","delivery":"Treat delivery as a supervised rollout on one instrument. Work covers converting existing scripts to MCP tools, writing experiment guidance for the model, connecting metadata and compute, and training staff to review agent actions. Ownership sits with the core facility manager. Governance checkpoints define which actions need confirmation. Acceptance criteria are proposals, such as task completion without operator correction on a test set, not results the Foundry reported."},"roleDetails":{"sales":{"customerProblem":"Advanced shared instruments need weeks of training and produce data faster than staff can process, limiting access for visiting researchers.","stakeholders":["Core facility director","Vice president for research","Research computing lead","Instrument staff scientists"],"discoveryQuestions":["Which instruments already have scripted automation?","How long does user training take today?","Where is experiment metadata stored, and can users search it?"],"valueHypothesis":"A plain-language front end over existing automation may let more users run standard techniques with less staff time, if validated.","potentialEngagement":"A pilot exposing one instrument's existing scripts through MCP with logging and staff review.","unsupportedClaims":["Measured training-time reductions","Error rates or safety of agent control","Results on instruments without existing automation"]},"engineering":{"fit":"Fits facilities with scriptable instruments and a metadata platform.","architectureIntegration":"MCP server wrapping pre-written instrument actions, a metadata catalog for recall, and HPC links for processing.","prerequisites":["Scriptable instrument control interface","Experiment metadata store","Approved commercial or local model service"],"constraints":["Vendor instrument software may lack open interfaces","No published reliability data in the article","Commercial model data terms"],"security":"Authenticate the MCP server, restrict actions to a reviewed list, and keep unpublished data on approved services.","proofOfValue":["Run established techniques and count operator corrections","Log and review every tool call","Compare processing time with and without HPC linkage"]},"delivery":{"work":"Convert scripts to MCP tools, write model guidance, connect metadata and compute, train staff and users.","dependencies":["Instrument vendor interfaces","Metadata platform","Model service approval"],"ownership":"Core facility manager with research computing support.","skillsAdoption":"Staff learn to maintain tools and correct agent behavior; users learn to state experiments clearly and check results.","governanceCheckpoints":["List of actions requiring human confirmation","Data classification before model access","Review after each new technique"],"acceptanceCriteria":["Agreed share of test tasks completed without correction (proposal)","All actions logged and attributable","No unapproved data leaves the facility"],"risks":["Agent confusing instrument actions","Over-reliance by inexperienced users","Model or vendor changes breaking tools"]}},"retrievedAt":"2026-10-29T22:01:28-05:00","enrichedAt":"2026-10-29T22:01:33-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Aims to let less-experienced visiting users run advanced techniques, potentially reducing weeks of hands-on training, though this is not measured.","procurementImplications":"Relies on a commercial model and in-house tooling; Crucible is being extended to other nanoscale research centers.","operatingModelImplications":"Staff scientists maintain the tool layer and teach the agent experiment rules; the agent becomes a front end to several systems.","peerReview":"unknown","sourceVerification":{"openedUrl":"https://foundry.lbl.gov/2026/10/29/new-virtual-assistant-streamlines-one-of-the-most-advanced-microscopes/","referenceExcerpt":"With TEM agent, I can just describe what I want in plain language, and it can string those steps together itself","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}