Agent Operating Layer

Saved workflow templates, reusable query packs, billing-aware agent guardrails, and policy bundles

Agent Operating Layer

The Agent Operating Layer provides infrastructure for running AI agents at scale against OMNI Datastream — with guardrails, templates, and billing controls.

Workflow Templates

Pre-built workflow templates that agents can execute with a single prompt:

Earnings Prep

"Prepare an earnings preview for {ticker} using OMNI Datastream"

Automatically calls: intelligence/earnings-preview, factors/decomposition, insiders, owners/13f

Compliance Check

"Run a compliance check on {company} using OMNI Datastream"

Automatically calls: events/enforcement, compensation, insiders, filings (8-K events)

Portfolio Review

"Review my portfolio using OMNI Datastream"

Automatically calls: portfolio/analyze, portfolio/stress-test, factors/dashboard

Query Packs

Reusable query configurations that bundle multiple API calls:

{
  "name": "weekly-insider-review",
  "queries": [
    {"endpoint": "/v1/insiders", "params": {"ticker": "{ticker}", "limit": 20}},
    {"endpoint": "/v1/owners/13f", "params": {"cik": "{manager_cik}"}},
    {"endpoint": "/v1/events/enforcement", "params": {"respondent": "{company_name}"}}
  ],
  "schedule": "weekly",
  "output": "markdown"
}

Billing-Aware Agent Guardrails

Budget Controls

Set per-agent spending limits to prevent bill shock:

# Set a $10/month budget for an agent
curl -X PUT "https://api.secapi.ai/v1/billing/budget" \
  -H "x-api-key: $OMNI_DATASTREAM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"spendCapCents": 1000, "softCapCents": 750}'

The agent receives Omni-Budget-Used headers on every response and gets warned at 75%, 90%, 95% thresholds.

Rate Limiting

Per-endpoint rate limits prevent agents from overwhelming the API:

Endpoint Family	Rate Limit
Entity resolve	600/min
Filing search	600/min
Section extract	180/min
Intelligence query	30/min
Semantic search	120/min

Agent Bootstrap

Create ephemeral, scoped credentials for agents:

# Sponsor creates a bootstrap token
curl -X POST "https://api.secapi.ai/v1/agent/bootstrap_tokens" \
  -H "Authorization: Bearer $BEARER_TOKEN" \
  -d '{"ttlSeconds": 3600, "scopes": ["read:sec"]}'

# Agent exchanges token for credentials
curl -X POST "https://api.secapi.ai/v1/agent/bootstrap" \
  -d '{"token": "BOOTSTRAP_TOKEN"}'

Policy Bundles by Vertical

Pre-configured policy sets for different customer types. For ready-to-run prompts that exercise these policy bundles end-to-end, see the agent prompt library — 50 v1 prompts across the same five personas.

Investment Manager Policy

• Full access to all SEC, factor, and macro endpoints
• Intelligence bundles enabled
• Portfolio analytics enabled
• Budget: usage-based (PAYG or Team plan)
• 12 ready-to-run prompts: /agents/prompt-library/investment-manager

Law Firm Policy

• Enforcement, filing search, entity resolution
• Event monitoring (8-K, restatements, auditor changes)
• Webhook delivery for real-time alerts
• Budget: per-query with approval threshold
• 12 ready-to-run prompts: /agents/prompt-library/law-firm

Insurance/Risk Policy

• Financial health indicators
• Enforcement history
• Compensation screening
• Restricted to read-only operations
• Budget: fixed monthly cap
• 6 ready-to-run prompts: /agents/prompt-library/insurance (with v2 expansion when AAER MCP wrap merges)

Monitoring

Track agent activity via the observability endpoint:

curl "https://api.secapi.ai/v1/observability" \
  -H "x-api-key: $OMNI_DATASTREAM_API_KEY"

Export detailed event logs:

curl "https://api.secapi.ai/v1/events/export?format=ndjson" \
  -H "x-api-key: $OMNI_DATASTREAM_API_KEY"