EXO-0_PROBE // AI_BIM_AGENT

PROBE Local AI BIM Assistant & MCP Console

Deploy PROBE AI locally to execute natural language commands, validate model configurations, and automate repetitive Revit tasks directly on your workstation.

Probe Chat ShellAI Host Connected

PROBE

>> PROBE STAGE_01 ONLINE. Secured tunnel to Autodesk Revit is open. Specify natural language commands to inspect parameters, run audits, or generate code. Local inference is active; 0% data leak risk.

PROBE

[
  {
    "id": 12345,
    "name": "Material-1"
  },
  {
    "id": 67890,
    "name": "Material-2"
  }
]

PROBE

In this case, we can see that the materials named "Material-1" and "Material-2" have standard names. However, if a material has a non-standard name like "Invalid Name", it would be reported in the output. Here is the MCP block to validate naming convention for all materials:

PROBE

{"command":"validate_naming_conventions", "params":{"elementType":"material"}}

MCP Bridge Server

RUNNING

API Endpoint:

Bridge Port:

node:

⚡ MCP Command Set166/166 active

Activate specific MCP bindings inside Autodesk Revit.

Context & Utility (4/4)

getactive_document_info

getselection_context

logaction_status

sayhello

Material & Naming APIs (7/7)

Temperature:0.2

Max Tokens:2048

MODEL_HEALTH_INTEGRITY

100.00%

MAX: 100.0%

MIN: 95.0%

SYS_STREAMING

TUNNEL: WORKSTATION_NODE (GOVERNED)LATENCY: 4ms

System Diagnostics

PROBE AI automates element requests. Model automation sequences run entirely locally on your workstation.

[System Status]ONLINE

[AI Engine Profile]WORKSTATION_NODE

[MCP Handshake]ACTIVE

AUTOMATION WORKFLOW

AI Generation Sequence

ASK

Provide natural language instructions via the chat shell:
"Create sheets for all levels"

PLAN

AI translates query to logical actions:
GetLevels() -> FilterNames()

EXECUTE

PROBE triggers native Revit sheet APIs:
Generate 4 title-block sheets

COMPLETE

Sheets are created in Revit layout:
✔ Sheet creation successful

MODEL INTERFACES

Use the AI That Fits Your Workflow

OpenAI Integration

Connect to GPT-4o cloud APIs for high-performance inference.

Claude Integration

Harness Claude 3.5 Sonnet for deep code generation and architectural compliance.

Ollama Local AI

Run offline models (Llama 3, Mistral) locally on your workstation.

Enterprise AI Server

Deploy a central private LLM gateway on your own secure enterprise server.