SemiOps AI Tools
SemiOps exposes AI tools for semiconductor facility and cleanroom operations. The tools help agents and product workflows review cleanroom state, ISO 14644 assessment history, pressure cascade, particle trends, SMT quality, fab energy, utilities, filter life, vibration, and decision-center evidence.
Use this page when an Agent workflow or backend integration needs a practical map of the SemiOps tools and the expected review path.
Tool layers
| Layer | Primary users | Access boundary | Output type |
|---|---|---|---|
/mcp/semiops/ MCP tools | Agent clients, copilots, workflow orchestrators | semiops.read scope | Cleanroom state, particle trends, pressure gradients, ISO status, SMT OEE, defect priorities, PUE, utilities, filter life, environmental analysis |
/ai/semiops/* AI Engine endpoints | SemiOps UI and backend workflows | Product service authentication and SemiOps permissions | ISO assessment, particle monitoring, pressure check, environmental prediction, soft-sensor output, filter prediction, health score, vibration assessment, energy and SMT analysis |
/api/v1/semiops/* module APIs | SemiOps UI and backend integration | semiops.view and semiops.edit permissions | Cleanrooms, SMT lines, utilities, maintenance views, compliance records, decision-center records |
/mcp/base/ MCP tools | Agent workflows that need cross-module context | base.read, base.compute.run, and selected base.action.write scopes | Documents, connectors, data quality, work records, action plans, supporting reports |
Use SemiOps MCP tools for cleanroom and electronics-manufacturing questions. Use base MCP tools when the answer needs documents, source data status, work orders, or cross-module evidence.
MCP endpoint and scope
| Endpoint | Scope | Use |
|---|---|---|
/mcp/semiops/ | semiops.read | Read cleanroom, SMT, utility, energy, maintenance, and compliance analysis output. |
/mcp/base/ | base.read | Add supporting documents, connector status, data quality, equipment records, and action-plan history. |
/mcp/base/ | base.compute.run | Generate reports or supporting analysis for a reviewed SemiOps workflow. |
SemiOps write actions stay in the product workflow under the configured semiops.edit permission and reviewer approval path.
MCP tool map
Cleanroom state and ISO review
| Tool | Primary job | Typical question |
|---|---|---|
get_cleanroom_status | Read cleanroom temperature, humidity, pressure, particle counts, and ISO status. | "Which cleanrooms need attention today?" |
get_iso_compliance | Read ISO 14644 compliance status and assessment history. | "What is the latest ISO assessment result?" |
get_particle_trend | Retrieve particle count trends by cleanroom, time window, and particle size. | "Did particle counts rise in this exposure area?" |
monitor_particles | Evaluate particle counts against ISO limits and produce alert details. | "Are current particle readings within the operating target?" |
get_pressure_gradient | Review pressure differential cascade status across cleanroom pairs. | "Is the pressure cascade maintained?" |
Environmental analysis and soft sensors
| Tool | Primary job | Typical question |
|---|---|---|
analyze_env_correlation | Analyze correlations between temperature, humidity, pressure, and particles. | "Which environmental parameters move together?" |
predict_env_trend | Predict short-term environmental trends for temperature, humidity, particles, or pressure. | "What is likely to happen over the next few hours?" |
run_soft_sensors | Estimate AMC, dew point, and HEPA or ULPA filter loading from available readings. | "What can we estimate when direct measurement is unavailable?" |
SMT quality and production
| Tool | Primary job | Typical question |
|---|---|---|
get_smt_oee | Read SMT line OEE with availability, performance, and quality breakdown. | "Which line has the weakest performance this week?" |
classify_smt_defects | Classify SMT defects, build Pareto-style priorities, and suggest root-cause review areas. | "Which defects should the quality team review first?" |
simulate_smt_bottleneck | Run or estimate an SMT bottleneck simulation. | "Which station limits throughput?" |
Energy, utilities, and maintenance
| Tool | Primary job | Typical question |
|---|---|---|
get_fab_pue | Read fab PUE and facility energy breakdown. | "How is facility energy performance trending?" |
forecast_fab_load | Forecast fab electrical load over the selected horizon. | "When may peak load occur?" |
optimize_chiller_cop | Compare chiller loading strategies and COP impact. | "Which chiller loading plan should engineering review?" |
get_utility_status | Read CDA, nitrogen, process cooling water, and ultra-pure water status. | "Are critical utilities stable?" |
get_filter_life | Estimate HEPA or ULPA cleanroom filter remaining life from pressure-drop trends. | "Which cleanroom filters need replacement planning?" |
AI Engine endpoint map
| Endpoint | Purpose | Inputs to prepare | Output to review |
|---|---|---|---|
/ai/semiops/iso-assess | Assess particle data against an ISO cleanroom target class. | Cleanroom ID, particle data by size, target class, sampling context. | Assessed class, pass state, per-size limits, means, and review notes. |
/ai/semiops/pressure-check | Evaluate pressure cascade readings. | Zone pairs, target pressure differential, actual reading, tolerance. | Normal, warning, reversed, or sensor-fault status per zone. |
/ai/semiops/particle-monitor | Evaluate particle readings against ISO limits. | Cleanroom ID, particle size, particle count, ISO class. | Per-size evaluation, alert details, and overall status. |
/ai/semiops/env-predict | Predict a short-term environmental parameter trend. | Parameter, recent values, timestamps, horizon, optional thresholds. | Predicted value, trend direction, confidence, and threshold timing. |
/ai/semiops/env-correlate | Compute correlations across time-aligned environmental parameters. | Temperature, humidity, pressure, particle, or other aligned series. | Pairwise correlations, strong relationships, and generated insights. |
/ai/semiops/soft-sensor | Estimate AMC, dew point, and filter loading. | Temperature, humidity, particles, air changes, cleanroom age, filter pressure drop, filter hours. | Estimated values, contributing factors, and confidence notes. |
/ai/semiops/filter-predict | Estimate filter life from pressure-drop history. | Cleanroom ID, filter pressure-drop history, threshold assumptions. | Remaining life estimate and replacement planning notes. |
/ai/semiops/health-score | Calculate equipment health for SemiOps assets. | Equipment age, alert count, maintenance state, sensor deviation. | Composite score, grade, and contributing dimensions. |
/ai/semiops/vibration-assess | Assess vibration against VC curve requirements. | Frequency-domain velocity measurements and target curve. | Violation details, severity, and review notes. |
/ai/semiops/pue and /ai/semiops/pue/trend | Calculate and summarize PUE. | Facility power, IT or process load, historical readings. | PUE value, trend, category, and change notes. |
/ai/semiops/load-forecast | Forecast fab electrical load. | Recent load history, horizon, profile assumptions. | Forecast values, peak periods, and demand-response review input. |
/ai/semiops/cop-optimize | Compare chiller loading plans. | Cooling demand, rated capacity, COP curves, ambient temperature, electricity cost. | Candidate loading plan, estimated energy impact, and review notes. |
/ai/semiops/defect-classify | Classify SMT defect records. | Defect records, line ID, product context, period. | Defect distribution, severity grouping, root-cause review areas. |
/ai/semiops/defect-dpmo | Calculate DPMO and approximate sigma level. | Defect count, unit count, opportunities per unit. | DPMO, sigma estimate, and quality review inputs. |
/ai/semiops/bottleneck-sim | Simulate or estimate SMT line throughput. | Station cycle times, uptime, buffers, simulation duration, board count. | Throughput, bottleneck station, utilization, and improvement candidates. |
Filter tool selection
Use get_filter_life for cleanroom HEPA and ULPA filter planning based on pressure-drop trends. Use Predictive Maintenance AI Tools for fleet or equipment component intelligence, recovery planning, and remaining-life workflows outside cleanroom filter operations.
Recommended tool sequences
Cleanroom daily review
get_cleanroom_status
-> monitor_particles for affected cleanrooms
-> get_pressure_gradient
-> get_iso_compliance
-> base tools for supporting documents or work records
Use this sequence for daily cleanroom status meetings and shift handover.
Environmental early warning
get_particle_trend
-> analyze_env_correlation
-> predict_env_trend
-> run_soft_sensors when direct measurements are unavailable
-> reviewed engineering summary
Use this sequence when environmental drift, contamination risk, or process sensitivity needs review.
SMT quality review
get_smt_oee
-> classify_smt_defects
-> simulate_smt_bottleneck
-> base document search for SOP or process notes
-> quality-team review package
Use this sequence when the team needs to connect line performance, defect concentration, and process constraints.
Energy and utilities review
get_fab_pue
-> get_utility_status
-> forecast_fab_load
-> optimize_chiller_cop for candidate review
-> generate_report through base MCP if a summary is needed
Use this sequence for facility energy and utility discussions. Confirm meter identity, load definition, operating mode, and engineering constraints before acting on candidate changes.
Maintenance and compliance handoff
get_filter_life
-> vibration assessment or health-score endpoint when equipment context is needed
-> get_iso_compliance
-> base tools for documents, work orders, and action-plan history
-> reviewer-approved maintenance plan
Use this sequence for filter replacement planning, cleanroom equipment review, and compliance evidence preparation.
Data requirements
SemiOps AI tools are strongest when the data package includes:
- stable tenant, fab, cleanroom, zone, line, equipment, utility, and source-system IDs;
- particle counts by size, sampling point, timestamp, and ISO target class;
- pressure differential readings with target values and tolerance;
- temperature, humidity, particle, pressure, AMC, dew point, vibration, utility, and energy readings with units;
- SMT production, OEE, defect, station, cycle-time, and product-context records;
- filter pressure-drop history, initial pressure drop, threshold pressure drop, operating hours, and replacement records;
- ISO assessment history, compliance evidence, SOPs, work orders, and maintenance schedules;
- DFS quality notes for timestamp alignment, stale values, missing intervals, unit conversion, and source mapping.
Use Create an AI Agent-Ready Dataset and Data Quality in DFS Lite before workflows that depend on fused cleanroom, utility, or SMT records.
Output review checklist
- Cleanroom, line, equipment, and utility identifiers are stable.
- Particle size, count units, pressure units, vibration units, power units, and time windows are visible.
- ISO target class, assessment date, and reviewer status are included when compliance output is shown.
- Environmental prediction and correlation outputs include source freshness and time alignment notes.
- Filter-life output includes pressure-drop history and replacement-threshold assumptions.
- SMT outputs show line, product context, defect period, and data coverage.
- Energy and chiller outputs show demand definition, COP assumptions, ambient conditions, and operating constraints.
- Recommendations are reviewed by the responsible facility, quality, manufacturing, or maintenance owner before field action.