Introducing
•
Trusys MCP
Trusys in your IDE, right now.
right now.
Connect Cursor, Claude Code, or Google Antigravity to the full Trusys platform via Model Context Protocol. Evaluate, red-team, monitor with TruPulse, and enforce TruGuard guardrails — without leaving your editor.
01
FUNCTIONAL
tru EVAL
Ready
02
RED-TEAM
tru SCOUT
Ready
03
MONITOR
tru PULSE
Ready
04
GUARDRAILS
tru GUARD
Ready
Natural Language Commands
Just ask. Your assistant handles the rest.
The MCP server maps natural language to the right Trusys API calls automatically.
Functional Evaluation
> Run an evaluation on my application
> Create a prompt library and eval localhost:3000
> Run a quality eval and show me the results
> Test my app against the prompt library
>
>
Run Security
Evaluation
Security red-team Evaluation
> Run a security eval on my app
> Red-team test the chatbot application
> Create a new security config with prompt-injection plugins
> Show me any vulnerabilities found
Monitoring
> Enable Pulse monitoring on my app
> Set up OpenTel for Trusys
> Generate traces for my project
> Enable monitoring with security plugins on traces
Guardrails
> Add guardrails to my app
> Enable TruGuard with prompt injection and PII checks
Set up input and output validators
Add TruGuard with content-safety and PII masking
> Set up input and output validators
Add TruGuard with content-safety and PII masking
> Add TruGuard with content-safety and PII masking
✣ Four products. One connection.
Everything Trusys can do,your assistant can now do
Trusys MCP exposes the full platform as natural-language tools your AI assistant calls on your behalf
Functional Evaluation
tru EVAL
skills/trusys-functional-eval
Run structured quality evaluations on your AI applications — before launch and on every release.
- Prompt libraries
Build and run reusable test suites against any application or LLM endpoint.
- Evaluations
Trigger TruEval runs from chat — compare versions, check regressions, get scores.
- Metric Catalogue
Choose from 100+ pre-built metrics or define your own quality dimensions.
Security Red-Teaming
tru SCOUT
skills/trusys-security-eval
Adversarially test your AI for prompt injection, jailbreaks, and policy violations — on demand, from chat.
- Security Configuration
Create or reuse configs with plugins from the Trusys red-team catalog.
- Automated red-team runs
The assistant spawns the Trusys CLI and runs full adversarial campaigns.
- Strategy Catalog
Basic, multi-turn, role-play, and encoded attack strategies out of the box.
Production Monitoring
tru PULSE
skills/trusys-pulse-monitoring
Enable OpenLIT tracing on your app and stream LLM traces to Trusys for continuous quality and security evaluation.
- OpenLIT integration
The assistant installs and wires openlit.init() into your app at startup.
- OTLP trace ingestion
Traces flow to otel.trusys.ai — evaluated against functional metrics and security plugins.
- Live dashboards
QA score, drift detection, language parity, and failure alerts in production.
Runtime Guardrails
tru GUARD
skills/trusys-guardrails
Configure validator rules and wire the trusys SDK into your code so every LLM call is scanned in real time.
- Validator Configuration
Input prompt-injection (fail), output content-safety (fail), PII (mask) — plus custom rules.
- SDK Wiring
The assistant adds TruGuard.init(), input_guard(), and output_guard() to your codebase.
- Guardrails Tab
Verify Enabled status in the Trusys dashboard and test with live adversarial input.
Agent Skills — optional, recommended
"If AI is powerful enough to replace human tasks, it's powerful enough to replace human-intensive governance. The future of AI assurance is conversational, automated, and intelligent—not manual spreadsheets and scripts."
"If AI is powerful enough to replace human tasks, it's powerful enough to replace human-intensive governance. The future of AI assurance is conversational, automated, and intelligent—not manual spreadsheets and scripts."
✓ skills/trusys-functional-eval
Why this exists
Your IDE already runs your code. It should run your AI governance too.
It should run your AI governance too.
You are shipping AI features. You are writing prompts, connecting tools, building agents. You are moving fast — and so are the risks: hallucinations, prompt injection, PII leakage, fairness drift.
The old way
❌ Switch context
❌ Open another tab
❌ Log into a governance platform
❌ Run a report
❌ Wait for someone else to find the problem
With Trusys MCP
✅ Stay in your IDE
✅ Ask your assistant
✅ Trusys MCP runs the call
✅ Get scores, traces, and verdicts
✅ Catch it before you ship
Trust & Security
Built for Enterprise Security Standard
Your AI evaluation infrastructure needs to be as secure and compliant as your production systems. TruEval is built with enterprise requirements in mind.
Enterprise-Grade Security
SOC 2 Type II certified. Your evaluation data is encrypted at rest and in transit.
Data Privacy First
Your prompts, datasets, and results stay in your environment. We never train on your data.
Global Compliance
GDPR, HIPAA, SOC 2 compliant. Built for regulated industries from day one.
Audit-Ready Documentation
Every evaluation is logged and versioned. Generate compliance reports on demand.
Flexible deployment options
Cloud-Hosted
Fully managed SaaS platform. Start evaluating in minutes.
On-Premise
Self-hosted in your data center. Maximum security and compliance.
Stop switching tabs.
Start Start governing inline.
iGet complete visibility into your production AI. Monitor quality, track costs, detect anomalies, and ship with confidence.
Questions about Trusys MCP?
Our team is here to help. Book a walkthrough to see how Trusys MCP fits your specific AI stack and IDE setup.
Book a Demo
Ready to dive in?
Check out the full user guide — endpoint reference, skill install, all four capability workflows, and troubleshooting in one place.
Start Free Trial
60 sec
To first connection
4
Products unlocked
3
AI IDEs supported
Benefits
Specifications
How-to
Contact Us
Learn More
Introducing
•
Trusys MCP
Trusys in your IDE,
right now.
Connect Cursor, Claude Code, or Google Antigravity to the full Trusys platform via Model Context Protocol. Evaluate, red-team, monitor with TruPulse, and enforce TruGuard guardrails — without leaving your editor.
01
FUNCTIONAL
tru EVAL
Ready
02
RED-TEAM
tru SCOUT
Ready
03
MONITOR
tru PULSE
Ready
04
GUARDRAILS
tru GUARD
Ready
Natural Language Commands
Just ask. Your assistant handles the rest.
The MCP server maps natural language to the right Trusys API calls automatically.
Functional Evaluation
> Run an evaluation on my application
> Create a prompt library and eval localhost:3000
> Run a quality eval and show me the results
> Test my app against the prompt library
>
>
Run Security
Evaluation
Security red-team Evaluation
> Run a security eval on my app
> Red-team test the chatbot application
> Create a new security config with prompt-injection plugins
> Show me any vulnerabilities found
Monitoring
> Enable Pulse monitoring on my app
> Set up OpenTel for Trusys
> Generate traces for my project
> Enable monitoring with security plugins on traces
Guardrails
> Add guardrails to my app
> Enable TruGuard with prompt injection and PII checks
> Set up input and output validators
> Add TruGuard with content-safety and PII masking
✣ Four products. One connection.
Everything Trusys can do,your assistant can now do
Trusys MCP exposes the full platform as natural-language tools your AI assistant calls on your behalf
tru EVAL
Functional Evaluation
skills/trusys-functional-eval
Run structured quality evaluations on your AI applications — before launch and on every release.
- Prompt libraries
Build and run reusable test suites against any application or LLM endpoint.
- Evaluations
Trigger TruEval runs from chat — compare versions, check regressions, get scores.
- Metric Catalogue
Choose from 100+ pre-built metrics or define your own quality dimensions.
tru SCOUT
Security Red-Teaming
skills/trusys-security-eval
Adversarially test your AI for prompt injection, jailbreaks, and policy violations — on demand, from chat.
- Security Configuration
Create or reuse configs with plugins from the Trusys red-team catalog.
- Automated red-team runs
The assistant spawns the Trusys CLI and runs full adversarial campaigns.
- Strategy Catalog
Basic, multi-turn, role-play, and encoded attack strategies out of the box.
tru PULSE
Production Monitoring
skills/trusys-pulse-monitoring
Enable OpenLIT tracing on your app and stream LLM traces to Trusys for continuous quality and security evaluation.
- OpenLIT integration
The assistant installs and wires openlit.init() into your app at startup.
- OTLP trace ingestion
Traces flow to otel.trusys.ai — evaluated against functional metrics and security plugins.
- Live dashboards
QA score, drift detection, language parity, and failure alerts in production.
tru GUARD
Runtime Guardrails
skills/trusys-guardrails
Configure validator rules and wire the trusys SDK into your code so every LLM call is scanned in real time.
- Validator Configuration
Input prompt-injection (fail), output content-safety (fail), PII (mask) — plus custom rules.
- SDK Wiring
The assistant adds TruGuard.init(), input_guard(), and output_guard() to your codebase.
- Guardrails Tab
Verify Enabled status in the Trusys dashboard and test with live adversarial input.
Agent Skills — optional, recommended
Install full workflow recipes for each capability
The base MCP connection gives your assistant access to all tools. Skills add detailed step-by-step recipes so the assistant never has to guess the right workflow.
✓ skills/trusys-functional-eval
✓ skills/trusys-security-eval
✓ skills/trusys-guardrails
✓ skills/trusys-pulse-monitoring
Why this exists
Your IDE already runs your code.
It should run your AI governance too.
You are shipping AI features. You are writing prompts, connecting tools, building agents. You are moving fast — and so are the risks: hallucinations, prompt injection, PII leakage, fairness drift.
The old way
❌ Switch context
❌ Open another tab
❌ Log into a governance platform
❌ Run a report
❌ Wait for someone else to find the problem
With Trusys MCP
✅ Stay in your IDE
✅ Ask your assistant
✅ Trusys MCP runs the call
✅ Get scores, traces, and verdicts
✅ Catch it before you ship
Trust & Security
Built for Enterprise Security Standard
Your AI evaluation infrastructure needs to be as secure and compliant as your production systems. TruEval is built with enterprise requirements in mind.
Enterprise-Grade Security
SOC 2 Type II certified. Your evaluation data is encrypted at rest and in transit.
Data Privacy First
Your prompts, datasets, and results stay in your environment. We never train on your data.
Global Compliance
GDPR, HIPAA, SOC 2 compliant. Built for regulated industries from day one.
Audit-Ready Documentation
Every evaluation is logged and versioned. Generate compliance reports on demand.
Flexible deployment options
Cloud-Hosted
Fully managed SaaS platform. Start evaluating in minutes.
On-Premise
Self-hosted in your data center. Maximum security and compliance.
Stop switching tabs.
Start Start governing inline.
One command connects Cursor, Claude Code, or Antigravity to the full Trusys platform. Eval, red-team, TruPulse monitoring, TruGuard guardrails — all from chat.
Questions about Trusys MCP?
Our team is here to help. Book a walkthrough to see how Trusys MCP fits your specific AI stack and IDE setup.
Book a Demo
Ready to dive in?
Check out the full user guide — endpoint reference, skill install, all four capability workflows, and troubleshooting in one place.
Start Free Trial
60 sec
To first connection
4
Products unlocked
3
AI IDEs supported