Check if an IC-CAP object exists at the specified path.

Create a new IC-CAP object without replacing an existing object unless overwrite=true. Plot guard: when creating a Plot onto a name a Transform/Input/Output already uses, the Plot is auto-suffixed '_plot' ('_plot2', ... on collision) and the response reports name_adjusted/final_name, so no name conflict is ever created.

Delete an IC-CAP object. Irreversible. Use object_type to disambiguate when Transform and Plot share the same name (e.g., object_type='Plot' to delete the Plot instead of the Transform).

Rename an IC-CAP object. Supported types: Model, Dut, Setup, Input, Output, Transform, Plot, Macro. NOT supported for: Par, MPar, DPar, TableVar, Main, InstrVar. Use object_type to disambiguate naming conflicts.

Copy an IC-CAP object without replacing an existing destination unless overwrite=true.

Get the full path name of an IC-CAP object.

Get the parent path of an IC-CAP object (excluding object name).

Get the value of an IC-CAP object. For parameters: supports minvalmax. For datasets: supports msb.

Set the value of an IC-CAP object. For parameters: supports minvalmax. For datasets: supports msb and force.

Get visible variable or parameter value from an object, searching the hierarchy.

Get a specific child object by name with optional type filtering.

Get all child objects of a specific type. WARNING: May hang when used on Transform paths; use setup_get_transforms instead.

Get a list of child object names of a specific type.

Delete all child objects of specified types.

Find a variable in the child variable table of an object.

Find a variable by name with scope control: 'upward_scope' (search hierarchy) or 'obj_owned' (direct children only).

Get parameter precision and working precision of an IC-CAP object.

Create a Model object. Model is the top-level container with circuit netlist, DUTs, parameters, and variables. Set overwrite=true explicitly to replace an existing Model.

Get the Model's SPICE circuit netlist.

Set the Model's SPICE circuit netlist.

Get all DUT names under a Model.

Get all Model variables (MVar) with their values.

Get all Model parameters (MPar) with their values.

Get all Macro names under a Model.

Create a DUT (Device Under Test) object. Set overwrite=true explicitly to replace an existing DUT.

Get all Setup names under a DUT.

Get all DUT variables (DVar) with their values.

Get all DUT parameters (DPar) with their values.

Create a Setup object. Set overwrite=true explicitly to replace an existing Setup. Ordering: Transforms may be created before or after setup_import_mdm — both orders are supported (live-verified 2026-09-02: derivative, Optimize, and PythonLocal transforms created after import, no hangs). Importing data before creating program transforms lets their UPDATE_AUTO auto-execution run against real data.

Run SPICE simulation on a Setup. The gateway executes it synchronously on IC-CAP's main thread; external MCP clients may submit it as a managed job.

Run measurement instruments on a Setup. Use run_async=true for a managed observable job.

Get all Input names in this Setup.

Get all Output names in this Setup.

Get all Transform names in this Setup.

Get all Plot names in this Setup.

Get all Setup variables (SVar) with their names and values.

Get the data length of the Setup's dataset.

Get complete Setup info: all input names, output names, transform names, plot names, and data length. Useful for verifying Setup structure.

Export Setup measurement or simulation data to an .mdm file. File is written directly on the server side, suitable for large datasets. file_path should be an absolute path on the server machine.

Import data from an .mdm file into a Setup using Import Create command. Automatically creates inputs/outputs structure from MDM file. Ordering: Transforms may be created before or after the import — both orders are supported and live-verified (2026-09-02: derivative/Optimize/PythonLocal created after import, no hangs, configuration read back effective). A bonus of importing first: program transforms whose Program starts with an UPDATE_AUTO line auto-execute on creation, so that execution runs against the already-imported data. Data is processed on the server side, suitable for large datasets.

Create and configure an Input object.

Mode (modedict) options:

• V: Voltage mode (+Node, -Node, Unit, Compliance)

• I: Current mode (ToNode, FromNode, Unit, Compliance)

• F: Frequency mode

• T: Time mode

• P: Parameter mode (ParamName, Unit)

• U: Voltage mode (+Node, -Node, Unit, Compliance)

• W: Power mode

Sweep Type (sweepdict) options:

• LIN: Linear sweep (SweepOrder, Start, Stop, # of Points)

• LOG: Logarithmic sweep (SweepOrder, Start, Stop, # of Points, Dec or Oct)

• LIST: List sweep (SweepOrder, # of Values, Values)

• CON: Constant (Value)

• SYNC: Synchronous sweep (Ratio, Offset, MasterSweep)

• LSYNC: List synchronous (MasterSweep)

Get a specific field value from an Input (e.g. 'Mode', 'Sweep Type', 'Start', 'Stop').

Set a specific field value for an Input.

Get all configurable field names for an Input.

Get the mode dictionary showing available modes and their required fields.

Get the sweep type dictionary showing available sweep types and their required fields.

Get Input dataset as dictionary. Supports msb (M/S/B) and imaginary flag.

Get Input dataset as DataFrame with optional matrix indexing.

Set Input data from a dictionary. String keys such as '("M", "11")' are converted to tuple keys for IC-CAP. Data format: {"('M', '11')": [val1, val2, ...]}.

Get the number of data points in the Input dataset.

Get the Input dataset type: 'M' (measured), 'S' (simulated), or 'B' (both).

Get measured data from an Input. Convenience method equivalent to get_data(msb='M').

Get simulated data from an Input. Convenience method equivalent to get_data(msb='S').

Get both measured and simulated data from an Input. Convenience method equivalent to get_data(msb='B').

Create and configure an Output object.

Mode (modedict) options:

• V: Voltage measurement (+Node, -Node, Unit)

• I: Current measurement (ToNode, FromNode, Unit)

• C: Capacitance measurement (HighNode, LowNode, Unit)

• G: Conductance measurement

• S: S-parameters (Port1, Port2, ACGround, Unit)

• H: H-parameters

• Z: Z-parameters

• Y: Y-parameters

• M: Multi-port S-parameters (PortType, NumPorts, ACGround, Unit)

Type: M=measured, S=simulated, B=both

Get a specific field value from an Output (e.g. 'Mode', 'Unit'). Only works on Output objects, NOT Plot objects.

Set a specific field value for an Output.

Get all configurable field names for an Output. Only works on Output objects, NOT Plot objects.

Get the mode dictionary for an Output.

Get Output dataset as dictionary.

Get Output dataset as DataFrame with optional matrix indexing.

Get the number of data points in the Output dataset.

Create a Model variable (MVar). Common uses: SIMULATOR, PARAMETER_PRECISION, WORKING_PRECISION.

Get a Model variable value.

Set a Model variable value. Pass scalar values as strings and arrays as JSON array strings such as '["a","b","c"]'.

Create a DUT variable (DVar). Common uses: DUT-level overrides of SIMULATOR, etc.

Get a DUT variable value.

Set a DUT variable value. Pass scalar values as strings and arrays as JSON array strings such as '["a","b"]'.

Create a Setup variable (SVar). Common uses: X_LOW, X_HIGH, sweep range parameters for optimizers.

Get a Setup variable value.

Set a Setup variable value. Pass scalar values as strings and arrays as JSON array strings such as '["a","b"]'.

Create a system variable (IVar) at Main level. var_name is the variable name only (no path prefix).

Get a system variable (IVar) value at Main level.

Set a system variable (IVar) at Main level. Pass scalar values as strings and arrays as JSON array strings such as '["a","b"]'.

Get the size of an ICCAP_ARRAY type variable. Only works on variables whose value is an ICCAP_ARRAY (set via mvar_set with a JSON array like '["a","b"]'). Returns null for scalar string values.

Get a Model parameter value. MPar objects are defined by the circuit netlist and cannot be created or deleted.

minvalmax options:

• Value: Current parameter value (default)

• Min/Max: Parameter range bounds

• Opt Min/Opt Max: Optimization range bounds

Set a Model parameter value. Use minvalmax to set bounds (Min, Max, Opt Min, Opt Max) instead of the current value.

Check if a Model parameter exists. Useful for verifying circuit netlist defines the expected parameter.

Get a DUT parameter value. Supports same minvalmax options as mpar_get.

Set a DUT parameter value.

Check if a DUT parameter exists.

Create DUT parameters by declaring them in a circuit page: IC-CAP has no direct DPar constructor, so parameters materialize from $dpar(NAME=value) declarations. circuit='model' (default, per IC-CAP docs the declaration appears in every DUT's parameter set) or circuit='dut' (DUT Test Circuit only). Prior declarations of the same name are replaced; after writing, each parameter is verified to exist and initialized to its value. Requires the DUT to exist; verify results with dpar_exists/dpar_get.

Batch-set parameter boundaries (Min, Max, Opt Min, Opt Max) for multiple MPar parameters in one call. Only works for MPar, not DPar.

Create a Transform and optionally set its Function type and Program content in one call.

ORDERING: Transforms may be created before or after setup_import_mdm — both orders are supported (live-verified 2026-09-02). Engineer PEL programs conventionally start with an UPDATE_AUTO line, which makes the transform auto-execute on creation; with data already imported, that execution runs against real data.

TRANSFORM TYPES AND THEIR SYNTAX REQUIREMENTS:

1. PythonLocal - Python script. Use transform_type='PythonLocal', program=''. Example: transform_create(transform_type='PythonLocal', program='print("hello")')

2. PEL/Program - IC-CAP PEL script. Use transform_type='PEL', program=''. WARNING: PEL uses return statement to output data. Direct assignment like beta = ic/ib causes parse error "Direct assignment to the data set beta is not allowed. Use the RETURN statement." Correct: return ic/ib. WRONG: beta = ic/ib. Example: transform_create(transform_type='PEL', program='return ic/ib')

3. Equation - Simple mathematical expression. Use transform_type='Equation' (or check funcdict for available types). Equations are evaluated directly; no return statement is needed. Note: IC-CAP requires lowercase 'equation' as function type; the handler auto-maps 'Equation' to 'equation'. Equation content is stored in the 'Input' field (NOT 'Program'); the handler handles this automatically. Example: transform_create(transform_type='Equation', program='ic/ib')

4. Built-in - IC-CAP built-in extraction functions (e.g. BJTDC_is_nf, BJTDC_fwd_gummel, PNCAPsimu, RBBcalc). Use transform_type=''. Function is set automatically. Built-in transforms require table fields to be filled (data references to other setups/transforms). Inspect the transform's table fields with tablevar_get and consult the installed IC-CAP documentation for required fields. Example: transform_create(transform_type='BJTDC_is_nf') Then fill table fields: tablevar_set('/path/to/transform/Forward VC', 'fearly/vc')

5. Optimize - Optimization transform. Use transform_type='Optimize'. Configure it with optimizer_setup or with transform_set_optim_field for individual fields. Example: transform_create(transform_type='Optimize') Then: optimizer_setup(transform_path='/path/to/optimize', ...)

Malformed transform types, programs, or fields can leave IC-CAP unresponsive. Always verify the created transform with transform_get_function.

Creating a Transform onto a name a Plot already uses creates a name conflict: the result then carries a warning (the Transform is never auto-renamed), and Plot-side field access afterwards requires object_type='Plot'. Prefer distinct names for new objects.

Consult the IC-CAP documentation installed with your licensed software for model- and version-specific transform fields.

Execute/Run a Transform. CRITICAL SAFETY CHECKS before executing:

1. Verify the Transform exists (use transform_get_function or setup_get_transforms)

2. Verify the Function type is correct (use transform_get_function)

3. Verify Program content is non-empty for PEL/PythonLocal types (use tablevar_get on transform_path+'/Program')

4. For built-in transforms, verify all required table fields are filled

An invalid function type, malformed PEL program, or missing table field can leave IC-CAP unresponsive and require a restart.

Get the Transform's Function type (e.g. 'PythonLocal', 'PEL', 'Optimize'). This reads the 'Function' child field.

Set the Transform's Function type (e.g. change from 'PythonLocal' to 'PEL'). This writes to the 'Function' child field. Note: For Optimize type, use transform_create with transform_type='Optimize' instead.

Get the dictionary of available Transform function types (funcdict). Returns list of function type names that this Transform can be set to.

List the table fields readable on a Transform for its current Function type. Ownership: '' (empty, grouping node) owns no fields; 'equation' -> Input; 'derivative' -> Input, Output; 'Optimize' -> Algorithm/Error plus Target/Simulated/X Min/X Max/Name/Min/Max rows (actual row counts probed read-only) and algorithm-filtered option fields; program types -> Program. 'Function' itself is always readable. Use before tablevar_get/tablevar_set to introspect instead of triggering ICCAP_FIELD_NOT_APPLICABLE.

Get a Transform dataset as a dictionary. Returns measured (M), simulated (S), or both (B) data. Set imag=true to include imaginary components.

Get a Transform dataset as a DataFrame. Supports matrix indexing with m and n. Set imag=true to include imaginary components.

Get the number of data points in the Transform's dataset. Returns 0 for an empty dataset.

Allocate data space for a Transform with specified size. Used to pre-allocate dataset before filling with data. Default creates both measured and simulated (B) data with dimension 1.

Set an Optimize Transform's table field value via STFV.

Configure IC-CAP optimizer Inputs, Parameters, and Options tables programmatically. Option field names are validated against the current algorithm; invalid names return success=false with an explanation.

Common fields:

• Function: Set to "Optimize" to enable the optimizer

• Algorithm: Optimization algorithm (full names or PEL short codes, auto-mapped). Unknown values are rejected before reaching IC-CAP (the native SetAlgorithm accepts anything without checking). On success the response carries current_value with current_value_source='this write': IC-CAP provides no way to read the algorithm back, so your own write is the authoritative record.

  • Full names: "Levenberg-Marquardt", "Random", "Hybrid", "Gradient", "Quasi-Newton", "Minimax", "Genetic", "Least Pth"

  • Short codes: "L", "R", "H", "B", "G", "N", "Q", "I"

• Error: Error type (full names or PEL short codes, auto-mapped). Same write-only caveat as Algorithm.

  • Full names: "Relative", "Absolute" Short codes: "R", "A"

• Inputs table (0-indexed rows):

  • "Target[i]": Target measured data, e.g. "ia.m"

  • "Simulated[i]": Simulated data, e.g. "ia.s"

  • "X Min[i]", "X Max[i]": X-axis range limits

• Parameters table (0-indexed rows):

  • "Name[i]": Parameter name, e.g. "MAIN.IS"

  • "Min[i]": Parameter minimum bound

  • "Max[i]": Parameter maximum bound

• Options table (algorithm-specific!):

  • Random (R): RMS error, Rand Iters, Rand Std Dev, Rand Reward, Rand Penalty, Rand Seed, Comb Filter

  • LM (L): RMS error, Max error, Max Evals, Param Tol, Function Tol, Comb Filter, Param Delta

  • Hybrid (H): RMS error, Max error, Rand Iters, Comb Filter, Param Tol, Function Tol, Param Delta

  • Minimax (N): RMS error, Max error, Max Evals, Param Tol, Function Tol, Max Iters, Comb Filter

  • All algorithms also support: X/Y Lower/Upper Bound

  • NOTE: Random uses 'Rand Iters' (NOT 'Max Iters'), LM uses 'Max Evals' (NOT 'Max Iters')

Read an Optimize Transform's table field value. Counterpart to transform_set_optim_field.

'Algorithm' and 'Error' are write-only in the IC-CAP API: the table fields do not exist and GTFV rejects them, so this tool returns readable=false with an explanation instead of an error. Track the value you set via transform_set_optim_field (its response echoes current_value); to inspect a transform configured elsewhere, open its GUI. Indexed fields (Target[0], Name[0], etc.) and options read reliably.

Configure an Optimize Transform in a single call. Replaces multiple transform_set_optim_field calls.

Sets Algorithm, Error type, Inputs table (Target/Simulated/X range), Parameters table (Name/Min/Max), and Options all at once. The Transform must already exist with Function='Optimize' (use transform_create with transform_type='Optimize' first).

CRITICAL: ALL optimizer parameters MUST have both min AND max bounds specified. Unbounded parameters cause optimization failures (e.g. IS/N compensation). The server serializes all IC-CAP API calls, so parallel calls are automatically queued. If any field fails to set, the call stops and reports which field failed.

Options are validated against the selected algorithm. Common options per algorithm:

• Random (R): RMS error, Rand Iters (NOT Max Iters!), Rand Std Dev, Rand Reward, Rand Penalty, Rand Seed, Comb Filter

• LM (L): RMS error, Max error, Max Evals, Param Tol, Function Tol, Comb Filter, Param Delta

• Hybrid (H): RMS error, Max error, Rand Iters, Comb Filter, Param Tol, Function Tol, Param Delta

• Minimax (N): RMS error, Max error, Max Evals, Param Tol, Function Tol, Max Iters, Comb Filter

• All algorithms also support: X/Y Lower/Upper Bound

Create and fully configure a Plot in one call, never creating a name conflict. Supports all IC-CAP report types: XY GRAPH, SMITH PLOT, POLAR PLOT, RI GRAPH, SCATTER PLOT, HISTOGRAM, CDF PLOT, MULTI PLOT. Field mapping is live-verified per type: XY GRAPH traces go to 'Y Data 0..7' + '# of Traces'; SMITH/POLAR/RI use 'Sweep Data' + 'Smith/Polar/RI Data 0..7'; SCATTER PLOT takes a single 'Y Data' (only y_data[0] is used); HISTOGRAM/CDF PLOT take a single 'Data-set' (pass data_set, or y_data[0]); MULTI PLOT takes sub-plot names in y_data -> 'Plot 0..N' + '# of Plots' (set 'Orientation' afterwards via tablevar_set). Conflict guard: when a Transform/Input/Output already uses the requested name, the Plot is auto-suffixed '_plot' ('_plot2', ... on collision) and the response reports name_adjusted/final_name. Reconfiguring an existing conflicted Plot (overwrite=true) writes fields through a rename-aside transaction.

Return a Plot's complete field configuration in one call, without repeated tablevar_get round-trips or missed fields. Field names are enumerated from the live object, so the set always matches the Report Type — XY GRAPH: Report Type, Header, Footer, X Data, Curve Data, # of Traces, X/Y/Y2 Axis Type, Y2 Data, Y Data 0..7; SMITH PLOT: Sweep Data + Smith Data 0..7; POLAR PLOT: Sweep Data + Polar Data 0..7; RI GRAPH: Sweep Data + RI Data 0..7; SCATTER PLOT: single X Data + Y Data + axis types; HISTOGRAM / CDF PLOT: Data-set; MULTI PLOT: # of Plots, # of GUIs, Orientation, Plots Per Row, Plot 0..N. When any non-Plot object (Transform, Input, or Output — e.g. an output current named 'id' with a plot also named 'id') shares the Plot's name, the read runs inside an automatic rename-aside transaction (annotated conflict=true, resolved_as='Plot', via='rename_txn') and the original name is restored afterwards. In/Out Graphically are GUI runtime state, not table fields, and are not accessible via the API.

Display a Plot in the IC-CAP GUI. Use run_async=true for a broker-managed job; direct gateway execution remains on the UI thread.

Capture an IC-CAP Plot window by name and save it to a local file for visual inspection.

Show absolute error (MAX/RMS) on Plot window. Equivalent to ShowAbsoluteError in IC-CAP GUI.

Show relative error (MAX/RMS) on Plot window. Equivalent to ShowRelativeError in IC-CAP GUI.

Create a Macro object. Optionally set its text content.

Execute/Run a Macro. Use run_async=true for a managed observable job.

Get the text content of a Macro.

Set the text content of a Macro.

List all Model names in the current IC-CAP session.

List all system variables (IVar) with their names and values.

Get a child object from Main by name. Optionally filter by type (Model, IVar).

Get a variable or parameter value visible from Main (searches hierarchy: variables, parameters, tablevars, datasets).

Save an IC-CAP object to disk. For Save As, arg is the destination filename: a bare name is written into the MCP workspace, an absolute path must satisfy the workspace path policy (see ICCAP_MCP_ALLOW_EXTERNAL_FILES); .mdl is appended when missing. The result is verified against the file system and reports file_exists/file_size - success means the file is actually on disk. Use locator='/' for the full project or a model path for one model.

Open a .mdl project file into the named model slot (IC-CAP's Open action reads the file over the target, destroying its previous contents). When the target model exists, a Save-As-No-Data checkpoint of its current contents is taken first and returned in the response. The file must exist and satisfy the workspace path policy (set ICCAP_MCP_ALLOW_EXTERNAL_FILES=1 to open from anywhere). Use this to restore a project after an IC-CAP restart.

Export a model's full structure to one JSON file in a single call: per DUT the DPAR values; per setup the inputs, outputs, setup variables, and every transform in order (PEL/Python program text, equation expressions, derivative sources with X/Y Data and Order, complete optimizer configuration with targets, parameter tables, and options). Replaces thousands of per-field tablevar_get calls for knowledge capture and build verification. file_path follows the workspace path policy (see ICCAP_MCP_WORKSPACE).

Get a table field value. Works on any IC-CAP object that has table fields: Transform ('transform_path/Program' for PEL content), Plot ('plot_path/Report Type', 'plot_path/X Data', 'plot_path/Y Data N'), etc. Use tablevar_exists first to verify the field exists. When a path has both a Transform and a Plot with the same name, reads default to the Transform and the result is annotated with conflict=true / resolved_as='Transform' (all common table fields are Transform-exclusive); pass object_type='Plot' to read the Plot side — when any non-Plot object (Transform, Input, or Output) shares the Plot's name, the read runs inside an automatic rename-aside transaction (conflict=true, resolved_as='Plot', via='rename_txn') that restores the original name afterwards. Reading a field the object does not own (e.g. Program on an empty-Function grouping transform, 'Plot 0' on an XY GRAPH) returns ICCAP_FIELD_NOT_APPLICABLE instead of touching the C API; use transform_get_fields / plot_get_fields to introspect first.

Set an existing table field; this tool cannot create fields. Use tablevar_exists first. On Transform/Plot name-conflict paths, writes default to the Transform (result annotated with conflict=true); pass object_type='Plot' to write the Plot side — on conflict paths this runs inside an automatic rename-aside transaction (necessary because a direct write on a conflict path silently targets the non-Plot object and reports success). Plot writes are validated against the plot's actual field set first: writing a field the current Report Type does not own (e.g. 'Plot 0' on an XY GRAPH, which STFV would accept while writing nothing) returns ICCAP_FIELD_NOT_APPLICABLE listing the available fields — never a false success. Transform-side writes are validated the same way against the Function type.

Check if a table field exists. Returns boolean. Use before tablevar_get/set to avoid errors on non-existent fields. On name-conflict paths (Plot sharing its name with a Transform/Input/Output) the untyped probe reflects the non-Plot side; pass object_type='Plot' to probe the Plot side through an automatic rename-aside transaction (annotated conflict=true, resolved_as='Plot').

Get an instrument option field value. Path is constructed as input_output_path/field_name.

Set an instrument option field value.

Check if an instrument option field exists.

Check the selected IC-CAP gateway and report its busy/session state plus feedback capabilities.

Read Error Message dialogs from the selected IC-CAP process and dismiss them by default so modal UI errors do not leave automation blocked.

Read the current IC-CAP Status window text and parse optimizer progress without modifying the project.

List live IC-CAP sessions without exposing authentication tokens.

Select the IC-CAP session used by subsequent calls in this MCP process.

Get the state, result, error, and latest progress of a managed IC-CAP job.

List recent managed IC-CAP jobs, newest first.

Read incremental IC-CAP status output for a managed job.

Request IC-CAP Stop Activity for a running job; failure is reported explicitly.