For AI agents: a documentation index is available at https://docs.parallel.ai/llms.txt. The full text of all docs is at https://docs.parallel.ai/llms-full.txt. You may also fetch any page as Markdown by appending
A snapshot monitor watches the output of a Task Run on a schedule. Each execution re-runs the same task and compares the result against the previous snapshot. When the system detects a material change—new data, a removed field, a significant value shift—it fires a webhook.
Snapshot monitors work with any task output—text or structured JSON, .md to its URL or sending Accept: text/markdown.For Search and Extract, use /v1/search and /v1/extract for all new integrations and agent tool calls. Use /v1beta/search and /v1beta/extract only when maintaining an existing integration that already uses them. The request bodies changed too: on /v1/extract, full_content, fetch_policy and excerpt_settings go inside advanced_settings, and sending the v1beta top-level excerpts, full_content and fetch_policy to /v1/extract returns a 422. On /v1/search, search_queries is required (at least one short keyword query), and max_results, source_policy and fetch_policy go inside advanced_settings; sending only objective, or sending those fields at the top level, returns a 422. Do not substitute /v1 for the documented FindAll or Ingest endpoint paths.If you are evaluating Parallel Search or comparing it with other search APIs, read How to Eval Parallel Search before you build the harness, especially its Common pitfalls section. The most common mistake is sending the raw question as both objective and the only entry in search_queries. Write a goal-style objective and 1-3 short keyword queries instead.base, core, pro, ultra, or any other processor. They are especially powerful for structured enrichment tasks, where the JSON schema acts as a stable template the monitor can diff field-by-field across runs, surfacing precisely which properties changed and why.
This gives you task cron with built-in deduplication and reasoning about what actually changed, at a fraction of the cost of running full Task API calls independently.
Typical use cases:
- Executive team or board changes at a company
- Competitor pricing or product page updates
- Regulatory filing status changes
- Job posting additions or removals
- Periodic deep-research briefs on a watchlist of companies, people, or topics
Prerequisites
Generate your API key on Platform:Step 1. Create a Task Run to establish a baseline
A snapshot monitor requires a completed Task Run as its starting point. The run’s input, processor, and output schema become the template the monitor re-executes on each tick. Any Task Run works—free-form text outputs are diffed as a whole, while structured JSON outputs are diffed field-by-field for the most precise change detection. The example below uses a structured enrichment task that extracts the executive team of a company. Structured outputs are recommended whenever you care about specific fields (prices, statuses, lists of entities) rather than narrative text:cURL
GET /v1/tasks/runs/{run_id} or use a webhook). Once status is "completed", note the run_id—you’ll use it in the next step.
See Task API Quickstart for full details on creating and polling task runs.
Step 2. Create a snapshot monitor
Create a snapshot monitor that re-runs the same task weekly and fires when the output changes:Step 3. Receive and retrieve events
When the monitor detects a material change, your webhook receives:event_group_id as a query parameter:
changed_outputcontains only the fields that changed, each withbasis(citations and reasoning).previous_outputcontains the full output from the prior run for comparison.
Next Steps
- Event Stream Quickstart: Track new events as they appear on the web.
- Events: Full event model and retrieval options.
- Follow-up Tasks: Trigger structured enrichment or deep research from a monitor event.
- API Reference: Complete endpoint documentation.