AgentOps Apify Builders Bundle

It's an Apify Actor, not a standalone SaaS product: you run it via the Apify Store, API, CLI, or MCP server, feed it a query and either sample mode or a list of approved public source URLs, and it writes rows to an Apify dataset. Per the vendor's own input/output schema, the six named capabilities (synthetic checks, schema QA, cost-anomaly detection, incident summaries, tool-firewall checks, repair suggestions) map one-to-one onto six other Actors already published by the same vendor under the same username: AI Agent Synthetic Monitor, Apify Dataset Quality Auditor v2, AI Agent Cost Sentinel, Zentra Agent Incident Scribe, AI Agent Tool Firewall, and Repair Bench v2. The bundle's own input schema calls these 'member Actors' and lets a buyer set a per-member timeout, which is consistent with the bundle orchestrating those Actors internally, though the vendor's source code is hidden so this can't be independently confirmed.

Per Apify's own actor API, this Actor has 2 total users (1 in the last 30/90 days), 88 total runs since it was created on 2026-07-08, and zero reviews or bookmarks. The last successful build was 2026-08-09. That is a genuinely young, lightly-used listing rather than an established one.

Pricing is pay-per-event at $0.029 per 'result-delivered' record ($29.00 per 1,000 results), confirmed on the Actor's own pricing page. There's no separate Apify platform-compute charge on top of that. This is a price increase from the $0.01 launch price set on 2026-07-08 to $0.029 on 2026-07-23 — roughly 3x within two weeks of launch — logged by Apify as 'Scheduled from Zentra ProductStudio pricing,' i.e. an automated schedule rather than a one-off manual change. Sample mode (the default) does not emit a billable event, so a buyer can inspect the output shape for free before running anything live.

The dataset schema declares 18 fields, but only 11 are marked required: evidence, recordId, sourceName, sourceUrl, dedupeKey, retrievedAt, score, scoreReasons, confidence, errors, and runSummary. The six fields that carry the actual business findings — riskScore, policyViolations, permissionFindings, paymentControls, supplyChainFindings, retentionFindings, and recommendedActions — are not in the required list, meaning any given row can omit them. The vendor states, in its own README, that a run which can't produce a schema-valid result fails outright rather than emitting a generic placeholder, and that missing values are left null or flagged rather than invented. That's a reasonable design if true, but with zero independent reviews it hasn't been tested by anyone but the vendor.

The example dataset row shown in the Actor's own documentation returns literal strings like 'Sample risk score for AgentOps for Apify Builders Bundle' and 'Sample recommended action for AgentOps for Apify Builders Bundle' — templated fixture placeholders, not example real findings. The vendor labels these explicitly as sample data, not claims about a live result, so this isn't a false claim on the vendor's part. But it means the free sample run does not let a buyer evaluate output quality; it only confirms the JSON shape.

The Actor is published by 'Zentra' (Apify username zentrafoundry, storefront at zentrafoundry.com). The zentrafoundry.com domain was registered 2026-07-01, one week before this Actor was created. Per Apify's own store API, the same username has 210 separate Actor listings published as of this check — spot-checking several (Temu Product Scraper, various permit/license monitors, the sibling 'member' Actors named above) shows the same pattern of 2 total users and zero reviews as this listing. Several of Zentra's Actor listings, including this one and the unrelated Franchise / New Location Expansion Monitor, use identically templated schema field descriptions ending in the phrase 'produced by the existing Product Studio identity' — consistent with the catalog being generated from a shared template rather than written per-product. There is no dedicated security, trust, or status page; the source code is marked hidden on Apify, so none of this can be independently audited beyond what the schema and README disclose.