Getting Started with Zipf

Create a Monitor

Describe what you want to watch in plain English. Zipf's AI agents will patrol your sources on the schedule you set and deliver signals when something changes.

1. Navigate to /monitors and click New Monitor
2. Describe your monitoring intent (e.g., “Track AI safety research papers from major labs”)
3. Set a cadence — every 6 hours, daily, or weekly
4. Optionally add notifications (email or Slack) and tags

Zipf assesses your intent for specificity and suggests improvements if it's too vague. You can also start from quick-start templates (Competitor Monitor, News Alert, Research Tracker, and more).

Your Feed

The Feed is your intelligence briefing. It aggregates results from all your monitors into a single prioritized view.

• Bottom line — A one-sentence executive summary of the period
• Findings by signal tier — Urgent, notable, and routine items grouped by importance
• Summary tiles — New findings count, briefing topics, active monitors

Filter by scope (All / Mine) and time window (6h / 24h / 3d / 7d) to focus on what matters.

Signal Scores

Every execution gets a signal score (0–100) that measures how actionable the results are. Scores are grouped into four tiers:

What drives the score: stop-condition match (+50), new results detected (+20), change rate (up to +15), recency (up to +15), and penalties for no-change or empty findings.

Monitor Detail

Click into any monitor to see its execution history. Each execution shows:

• • Signal dot with score and tier color
• • AI-generated headline summarizing what changed
• • Key findings with source citations and badges (New Updated Context)
• • Change chips showing net new, dropped, and change rate

AI Edits

Refine a monitor without recreating it. Type a plain-English instruction like “Focus more on pricing changes and exclude opinion pieces” and the system adjusts the monitor's behavior. Every edit is recorded in the change timeline.

Chronicle

Individual executions show what happened now. Chronicle answers the bigger question: what's been happening over time?

• • Narrative arc — Executive summary of the monitor's history and current phase
• • Signal curve — Time-series chart of signal scores across executions
• • Topic threads — Recurring themes with trajectory (rising, stable, fading)
• • Entity tracks — Companies, people, and products tracked across findings
• • Coverage assessment — Source diversity and potential blind spots

Access Chronicle from any monitor's detail page. Select a time window: 1 day, 7 days, 30 days, 90 days, or all time.

Tags & Collections

Tag monitors to group them (e.g., #competitive, #ai-research). Collection Chronicle provides cross-monitor analysis for a tag group — a unified briefing across all monitors in that category.

Notifications

Get alerts where you work:

• • Per-monitor — Email or Slack webhook when a patrol finds something above your suppression threshold
• • Suppression threshold — Critical (80+), Important (50+), Balanced (20+, default), or All
• • Global email digest — Scheduled briefing to your inbox (every 6 hours, daily, or weekly)

Enable the digest at Dashboard → Account → Notifications. It's the single best way to stay informed without checking the dashboard.

Quick Setup

Any MCP-compatible AI agent can create, manage, and consume Zipf monitors — Claude Code, Codex, Gemini CLI, Cursor, or your own agents. OAuth handles authentication. Open protocol, not locked to one platform.

codex mcp add zipf --url https://api.zipf.ai/mcp

What You Can Do

Example Prompts

Paste these into your AI assistant after connecting Zipf:

Get Your API Key

Create a token at Dashboard → Tokens, then set it as an environment variable:

export ZIPF_API_KEY="wvr_your_token_here"

Quick Reference

Example: Create a Monitor

curl -X POST https://api.zipf.ai/v1/workflows \
  -H "Authorization: Bearer $ZIPF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "AI Safety Research Monitor",
    "mode": "simple",
    "workflow_type": "search",
    "operation_config": {
      "query": "AI safety research breakthroughs 2026",
      "max_results": 20
    },
    "stop_condition": {
      "type": "natural_language",
      "description": "Stop when a major AI safety breakthrough is announced",
      "confidence_threshold": 0.8
    },
    "interval": "1 day",
    "max_executions": 100
  }'

Example: Search

curl -X POST https://api.zipf.ai/v1/searches \
  -H "Authorization: Bearer $ZIPF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "AI infrastructure startups 2026",
    "max_results": 10
  }'

Full API Documentation