Private Beta • Invite Only

See what your AI decided, why, and what happened next

Control-plane observability for agentic workflows. Trace decisions, surface risk signals, enforce policy checks, and maintain audit-ready logs — before your AI takes action.

Private beta — open to teams building agentic AI systems

AI Logbook Console
support-agent-run
completed
2.4s total
StepStatusLatency
analyze_ticketmodel_call
OK
820ms
lookup_customertool_use
OK
340ms
fetch_policy_docsretrieval
OK
180ms
generate_resolutionmodel_call
OK
1.1s
ticket_resolution
requires_review
refund $250|conf: 0.87|financial_action
Steps:4
Decisions:1
Risk flags:1
trace_7f8a2b...

Private Beta

SOC 2 Ready
API-First
Secure by Design
Built for Agentic AI
Developer Experience

Instrument Your AI in Minutes

Wrap your existing OpenAI, Anthropic, or any LLM calls. No proxies, no lock-in.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
from ailogbook import AILogsClient
import openai

client = AILogsClient(
    base_url="https://api.ailogs.io",
    api_key="YOUR_KEY_ID:YOUR_SECRET"
)

with client.trace(trace_key="support-agent-run") as t:
    # Record each step of your agent
    t.step(step_type="model_call", name="analyze_ticket",
           model_name="gpt-4o")
    response = openai.chat.completions.create(
        model="gpt-4o",
        messages=[{"role": "user", "content": ticket}]
    )

    # Record the decision with risk scoring
    t.decision(
        decision_type="action_selection",
        decision_label="ticket_resolution",
        decision_value={"action": "refund", "amount": 250},
        confidence=0.87,
        action_status="requires_review",
        risk_flags=["financial_action"],
    )

How It Works

Install & Instrument

pip install ailogbook. Wrap your AI calls in client.trace(). Record decisions and policy checks with a few lines of code.

Observe in Real Time

Every trace, decision, and risk signal appears in your dashboard instantly. Filter by status, risk level, or time range.

Review & Export

Approve or reject flagged decisions. Export audit trails for compliance. Get webhook alerts for policy violations.

Key Features

Everything you need to observe, govern, and audit your AI systems.

Execution Traces
Wrap any AI workflow in client.trace(). Every model call, tool use, and retrieval step is captured with latency, tokens, and metadata.
Decision Records
Record what your AI decided, why, with confidence scores and risk flags. Distinguish proposed vs. approved vs. executed actions.
Policy Evaluations
Attach guardrail checks to any decision. Know which policies passed, failed, or triggered human review.
Human Review Gates
Flag high-risk decisions for human approval. Approve or reject from the dashboard with a full audit trail.
Webhook Alerts
Get notified instantly when a policy fails, a decision is blocked, or a trace completes with risk flags.
Compliance Exports
Export decisions, traces, and policy evaluations as CSV or JSON for audits.

Security & Tenant Isolation

AI Logbook isolates tenant data by default and encrypts everything in transit and at rest.

Encryption in Transit & at Rest
All data encrypted with TLS 1.3 in transit and AES-256 at rest.
Tenant Isolation
All data is scoped by tenant. Queries, exports, and API keys are strictly isolated.
Audit Logging
Every access and modification is logged with user, timestamp, and action for full auditability.
Plaintext-Off by Default
Sensitive prompt and response content is not stored unless you explicitly enable ALLOW_STORE_PLAINTEXT.

Private Beta — Now Open

We're building AI Logbook with early design partners who need decision audit trails and risk observability for agentic workflows. Request an invite to shape the product.

Request Access