One MCP server. Seventeen tools. 31+ resources. Connect once and your AI assistant becomes an automotive engineering expert -- inside Claude Code, Cursor, or Windsurf.
# Determine ASIL for an electric power steering failure > What ASIL for loss of EPS assist at highway speed? Calling tracely.determine_asil severity: "S3" # Life-threatening exposure: "E4" # Highway driving controllability: "C3" # Difficult to control Result: ASIL D Per ISO 26262-3 Table 4, S3/E4/C3 yields ASIL D. This requires: -- Hardware metrics SPFM >= 99%, LFM >= 90% -- Systematic capability SC 4 -- Dependent failure analysis (Part 9, Clause 7) -- All verification methods from Part 4, Table 9
Works with Claude Code, Cursor, Windsurf, and any MCP-compatible client. No SDK. No training data. Just one config entry.
Request access and we will send you a key within 24 hours. Free tier: 50 calls/day, no credit card.
Add this to .mcp.json in your project root (or ~/.claude/.mcp.json for global access):
{ "mcpServers": { "tracely": { "type": "url", "url": "https://tracely-mcp.autospec.workers.dev/mcp", "headers": { "Authorization": "Bearer YOUR_API_KEY" } } } }
Download tracely.md and place it at .claude/commands/tracely.md in your project. This gives you the /tracely slash command with usage examples and workflows.
Open Claude Code and ask anything about ASPICE or ISO 26262. The tools are called automatically based on your question.
> Calculate ASIL for a battery thermal runaway at parking > Draft a requirement for CAN timeout detection at ASIL C > Review this requirement: "The system shall be safe" > What work products do I need for SWE.3 at level 2? > Check my ASIL D decomposition into B + C > Assess our SWE.1 compliance against this evidence... > Review my software architecture for ASPICE SWE.2 compliance > Check traceability from system reqs to detailed design > Evaluate resource allocation: CPU 72%, RAM 85%, CAN 65%
Each tool takes typed inputs and returns actionable, referenceable outputs. Your AI assistant calls them automatically.
ASIL determination, decomposition, process compliance, requirements review and drafting, architecture assessment, interface review, dynamic behavior evaluation, detailed design, quality metrics, resource allocation, traceability checks, and expert consulting from a seasoned ASPICE change agent.
Every ASPICE 4.0 process area, ISO 26262 parts 1-12, cross-standard mappings, capability level definitions. Browsable and searchable.
Pre-built workflows for requirements review and architecture assessment. More coming: safety analysis, ASPICE migration, audit prep.
Full ASPICE 4.0 (December 2023) knowledge base. Not the outdated 3.1. Includes all new process areas, outcomes, and base practices.
Structured inputs, deterministic outputs. Each tool references specific ASPICE base practices and ISO 26262 clauses.
// Input { "severity": "S3", "exposure": "E4", "controllability": "C3" } // Output { "asil": "D", "reference": "ISO 26262-3:2018, Table 4" }
// Input { "text": "The system shall be fast", "type": "software" } // Output: score 2/10 { "issues": ["Not quantifiable", "Missing acceptance criterion"] }
// Input { "processId": "SWE.1", "evidenceDescription": "..." } // Output: findings per base practice { "gaps": ["BP5: No bidirectional trace"], "remediation": [...] }
// Input { "intent": "monitor tire pressure at 1 Hz, warn below 1.8 bar", "type": "safety", "asil": "B" } // Output: draft + quality checklist + template { "draft": "The TPMS ECU shall...", "checklist": [7 criteria], "asilGuidance": [...] }
// Input { "requirementText": "...", "asil": "C" } // Output { "methods": [ { "method": "Walk-through", "recommendation": "Highly recommended" }] }
// Input { "parentAsil": "D", "childAsils": ["B", "C"] } // Output { "valid": true, "independence": "Required", "reference": "ISO 26262-9, Clause 5" }
// Input { "processId": "SWE.3", "capabilityLevel": 2 } // Output: base practices + work products { "basePractices": ["BP1..BP7"], "workProducts": ["17-08", ...] }
// Input { "description": "EPS system: 3 SWCs, AUTOSAR Classic, CAN/LIN...", "level": "software", "asil": "D" } // Output: score + ASPICE findings { "score": 60, "findings": [ { "ref": "SWE.2.BP1", ... }] }
// Input { "description": "CAN msg 0x120, 8 bytes, 10ms cycle, E2E Profile 1", "level": "software" } // Output: interface quality findings { "score": 75, "findings": [ "Missing error handling spec"] }
// Input { "diagramType": "state_machine", "description": "Init->Running-> Degraded->SafeState", "asil": "C" } // Output: dynamic behavior findings { "score": 70, "ref": "SWE.2.BP4" }
// Input { "description": "PID controller: Kp, Ki, Kd params, 1ms cycle, anti-windup, MISRA C:2023", "asil": "D" } // Output: detailed design findings { "score": 82, "ref": "SWE.3.BP1-BP7" }
// Input { "description": "12 modules, avg cyclomatic complexity 8, 15% coupling", "level": "software" } // Output: 6 metric assessments { "metrics": [{ "name": "modularity", "status": "pass" }, ...] }
// Input { "description": "CPU 72% peak, RAM 85% used, CAN bus 65% load, 10ms deadline met" } // Output: resource budget analysis { "resources": [{ "type": "cpu", "status": "warning" }, ...] }
// Input { "description": "DOORS: SysReq-> SwReq->Design->Code, bidirectional", "asil": "D" } // Output: traceability assessment { "score": 100, "links": ["all traced"] }
// Input { "topic": "aspice_implementation", "question": "Our team spends 60% of time on docs. How do we fix this?" } // Output: expert perspective + source traces { "position": "ASPICE is not the bottleneck. Manual compliance is.", "recommendations": [...], "anti_patterns": [...] }
// Input { "topic": "ai_agents_engineering" } // Output: full framework + 8-step approach { "principles": ["Engineers build. AI documents. Compliance happens in the background.", ...], "agent_types": [7 agents] }
// Input { "topic": "config_management" } // Output: what NOT to do { "anti_patterns": [ "Excel-based config tracking", "Manual baseline assembly", "Chasing engineers for baselines after releases are done"] }
Full knowledge base on every tier. Higher tiers unlock more calls and priority support.
How tracely changes the way you work with ASPICE and ISO 26262.
Automotive safety engineers spend hours cross-referencing ISO 26262 Part 3 tables for ASIL calculations. With tracely, ask your AI assistant and get the result with full standard references instantly. No more flipping through 500-page PDFs.
Before tracely: Open PDF, find Part 3 Table 4, cross-reference severity/exposure/controllability manually, document the rationale. 15-30 minutes per hazardous event.
With tracely: "What ASIL for loss of braking at highway speed?" -- answer in 10 seconds with full traceability to ISO 26262-3:2018.
ASPICE assessments cost $15K-$50K per engagement. tracely lets you run continuous self-assessments against every process area, identifying gaps before the assessor arrives.
Feed your current evidence (documents, review logs, test reports) into assess_process_compliance and get a structured gap analysis per base practice. Know exactly where you stand on SWE.1 through SWE.6, SYS.1 through SYS.5, and every other process area -- before paying for an external assessment.
Model Context Protocol (MCP) is how AI coding tools like Claude Code access external knowledge. Instead of hoping the AI "knows" about ASPICE, MCP gives it structured, authoritative tools to call.
Think of it like this: without MCP, asking Claude about ASIL calculations is like asking a smart person who read a textbook once. With tracely via MCP, it is like having a certified ASPICE assessor sitting next to you -- with the full standard indexed and tool-accessible.
MCP for automotive engineeringWe onboard users manually during early access. Tell us about your use case and we will send your API key within 24 hours.
Free tier: 50 calls/day. No credit card required.