🐯 TealTiger v1.4.0 Released — Observe Mode & Governance Dashboard

[nagasatish007]

🐯 TealTiger v1.4.0 Released — Observe Mode & Governance Dashboard

We're excited to ship TealTiger v1.4.0 — the easiest way to get visibility into your AI agents.

observe() — One Line, Zero Config

import { observe, freeze } from 'tealtiger';
const client = observe(new OpenAI());
console.log(client.getCost()); // { totalCost: 0.0023, requestCount: 1 }

from tealtiger.observe import observe, freeze
client = observe(OpenAI())
print(client.get_cost())  # ObserveCostSummary(total_cost=0.0023, ...)

Wrap any of 12 supported LLM providers. Get cost tracking, audit logging, PII detection, and behavioral baselines instantly. Under 5ms overhead per call.

What's New

Feature	Description
observe(client)	Zero-config proxy wrapping for 12 LLM providers
Automatic Cost Tracking	Per-request, per-session, per-agent accumulation
Behavioral Baseline	P50/P95/P99 latency, token distribution, cost pattern profiling
PII Detection (REPORT_ONLY)	Passive scanning without blocking — visibility before enforcement
freeze() / unfreeze()	Instant kill switch to halt any agent, zero policy required
Governance Dashboard	Redesigned with light theme, color-coded KPIs, defense pipeline, canary alerts, agent matrix, protocol governance
Widget Error Isolation	Each dashboard panel renders independently — one failure never cascades

Governance Dashboard

Real-time visibility into your AI agent fleet — security posture, cost governance, and behavioral alerts in one view.

What you see at a glance:

• KPI Row — Total requests, cost, governance denials, budget consumption with color-coded indicators
• Cost Velocity & Budget Forecast — Burn rate trends and exhaustion projection
• Defense Pipeline — 3-stage security evaluation flow with short-circuit rates and latency per stage
• Canary Alerts — Behavioral drift detection with agent freeze status and deviation percentages
• Agent Matrix — Fleet status table (active/idle/frozen) with per-agent request metrics
• Cost Savings — Optimization recommendations ranked by impact
• Model Routing — Source-to-target routing with per-request savings
• Protocol Governance — ENFORCE/MONITOR/REPORT_ONLY policy cards with denial counts

Run locally: cd dashboard/api && npm run dev then cd dashboard/web && npm run dev

Progressive Disclosure

Level	Entry Point	What You Get
0	observe(client)	Cost, audit, PII detection, baseline, kill switch
1	+ guardrails config	Prompt injection, content moderation, secrets
2	+ TealEngine policies	ENFORCE / MONITOR / REPORT_ONLY per rule
3	+ TealFlow workflows	Org-level governance inheritance

12 Supported Providers

OpenAI · Anthropic · Google Gemini · AWS Bedrock · Azure OpenAI · Cohere · Mistral · DeepSeek · Groq · Together AI · HuggingFace TGI · xAI (Grok)

Install

npm install tealtiger@1.4.0
pip install tealtiger==1.4.0

Links

• GitHub: https://github.com/agentguard-ai/tealtiger
• Release Notes: https://github.com/agentguard-ai/tealtiger/releases/tag/v1.4.0
• npm: https://www.npmjs.com/package/tealtiger
• PyPI: https://pypi.org/project/tealtiger/
• Discord: https://discord.gg/X2ePf8QAj

What's Next (v1.5)

• Multi-tenancy with data isolation
• RBAC (Owner, Admin, Policy Author, Viewer, Auditor)
• SSO via SAML 2.0 / OIDC
• SIEM export (Splunk, Elastic, Sentinel, Datadog)
• Policy staging and canary deployments

---

Questions? Open a Discussion or join us on Discord.

Thanks to everyone who contributed feedback and tested early builds. 🧶