Your measurement database holds the answers. Getting to them should not require learning a query language. The ASAM ODS MCP Server connects AI assistants directly to your test data. Engineers ask questions in natural language. The server translates to HQL, validates against your application model, executes the query, and returns results. No syntax to memorize. No schema to study. Just ask what you need.
Connect Claude Desktop, ChatGPT, Claire, or any MCP compatible assistant to your ASAM ODS infrastructure. Your data stays where it is. The AI brings the interface.
Built-in browser-based chat for teams who want natural language access without installing anything. Ask questions, view results, explore data visually. Session memory keeps context across follow-up questions. Plotly charts for quick visualization. Share access across your organization through a URL.
Connect the MCP Server to Claude Desktop for a conversational AI assistant with direct database access. Query test data alongside other tasks. Export results into documents or code. Ideal for engineers who already use Claude for analysis and documentation work.
The MCP protocol is an open standard. Build natural language access into your own tools. Connect from Jupyter notebooks, internal applications, or automated pipelines. ChatGPT with MCP support, custom agents, or any system that speaks MCP. Your interface, our translation layer.
Ask about your test data in plain English. Find measurements by project, date, status, or any attribute in your application model. Filter and combine criteria naturally. Results come back as structured data your AI assistant can summarize, analyze, or export.
Go beyond single queries. Compare channel values across multiple test runs. Calculate statistics like min, max, average, and standard deviation. Ask follow-up questions to refine your analysis. Build toward insights through conversation.
Generate plots directly from your queries. View time series data as charts. Spot trends and anomalies visually. Export results as tables, charts, or raw data for further processing in your preferred tools.
Every application model is different. Your entity names, attribute structures, and relationships reflect your engineering domain. The server learns your schema dynamically and translates natural language queries into HQL that matches your specific model. No manual mapping. No generic assumptions. Ask in your terminology, get results from your data.
Translation alone is not enough. Every generated query is validated against your application model before execution. Element names, attributes, and relationships are verified. When something does not match, the server attempts automatic correction for common issues like case mismatches or typos. When correction is not possible, you receive specific guidance, not cryptic errors.
Your organization controls the AI access. Use your corporate Anthropic or OpenAI API keys, managed through secure environment variables. Credentials never appear in logs or query results. Centralized key management means IT controls access, monitors usage, and enforces policies. The AI works for you, under your governance.
Accessing ASAM ODS data requires specialized knowledge. Engineers must learn HQL syntax, understand the application model schema, and know which entities contain the data they need.
For occasional users, this is a barrier. They wait for database specialists to run queries, or they give up and work with whatever exports they already have. Critical data stays locked in the system because the interface is too complex.
Even experienced users spend time constructing queries, checking syntax, and debugging failed attempts. Time spent on mechanics, not analysis.
With MCP servers. you simply ask what you want to know. "Show me all measurements from project Alpha last month." "What channels are available in test run 4523" "Compare brake_pressure acroos test x and y".
The MCP Server handles translation. It parses your question, generates valid HQL, validates element and attribute names against your specific application model, and executes the query. You see results, not query syntax.
Non technical users can access data immediately. Technical users save time on routine queries. Everyone benefits from faster access to test data.
Find relevant measurements without learning query syntax. Check what data exists for a project. Locate specific test runs by date, vehicle, or configuration. Trigger standard analyses without navigating Merlin. Get quick answers without waiting for database support. Spend time on engineering, not data retrieval.
Explore unfamiliar datasets through conversation. Understand application model structure by asking questions. Prototype queries in natural language before formalizing in scripts. Export data in the format you need without configuration dialogs. Focus on insights, not interface mechanics.
Access test status and progress without technical assistance. Ask about measurements completed, analyses pending, or data quality issues. Get visibility into test programs without becoming a database expert. Make decisions based on current information, not last week's report.
Claire is built on pyHQL, HighQSoft's Python library for ASAM ODS. This is not experimental integration. pyHQL has been in production at major OEMs for years, handling both metadata queries and time series retrieval.
The same foundation that powers BMW, Ford, Volkswagen, and Bosch test data pipelines now powers natural language access. Production proven, not demo ware.
HighQSoft has been solving test data management problems since the 1990s. We understand application models, measurement workflows, and analysis automation because we have built production systems for the world's leading automotive companies.
Claire inherits this expertise. Natural language access that understands engineering context, not generic AI chat.
HighQSoft GmbH
Black-und-Decker-Straße 17b
D-65510 Idstein
