fetch
ChatGPTsearch tool.By SiteTrax.io
SiteTrax.io is a logistics operations analytics app for authorized teams. It can retrieve a request-scoped projection of the authenticated user's latest published SiteTrax.io Intelligence profile when saved role, priorities, user-confirmed facts, operational context, or project scope would materially improve customer-specific analysis. Search shared projects and yards, inspect container and supply chain equipment detections, review camera health and anomalies, analyze videos and images, retrieve audit feedback, and run server-side analytics for status, volume, activity, and sequence metrics. Look up SiteTrax.io logistics definitions and fetch research-ready reports including labor efficiency, revenue, and cost reduction opportunities. Create confirmed email alerts and scheduled digests for matching detections.
search tool.fetch with a returned id for full text. Args: query: Natural-language search query. limit: Maximum number of fetchable results to return.text (exact) and/or where (sitetrax_query_assets grammar); omit/[] projects = all shared. direction = ingate/outgate from asset_heading+l2risout; for gate projects/cameras scope by those instead. repeat=false completes after one match; true = de-bounced standing watch. texts + aware ISO expect_by: one email when all arrive, or at the deadline listing what is missing. prompt: AI writes the email from the data. expires_in_days=0 = 30 days. Mail goes only to the user's verified address. Requires the user's prior explicit confirmation.deliveries supports where/order_by/limit/offset/fields/expected_result_version. Delivery records are reauthorized against current Shares and omit contact data; event items include detection lineage, while digest items include their aggregate window and never invent a detection.gap operator. Optional bounded records selects native duplicate_groups, api_records, or low_days using where/order_by/limit/offset/fields and expected_result_version. Use returned available_record_types and never recalculate the finding from these details. Omit projects/pass [] for all shared projects; date inputs define the echoed resolved window.asset_id, including its raw capture payload and newest-first persisted submitted-feedback page. feedback_limit/feedback_offset page feedback; redacted detections expose media_redacted=true instead of media URLs.text as stored URL strings plus metadata. Date inputs or hours_back bound the search; limit=1 is the default. The server never fetches/base64-encodes media; legacy as_base64/max_bytes are ignored. Redacted records expose media_redacted=true instead of URLs.text and/or shared projects. text is an equipment identifier, not a numeric detection ID. verbose expands stored metadata but still omits the raw capture payload; use sitetrax_asset_detail for exact numeric-ID detail.text, including detection/project lineage and stored metadata/thumbnail URLs. Date inputs or hours_back bound the search; limit=1 is the default. No media is fetched and no playback URL is minted; use returned portal IDs. Redacted clips expose media_redacted=true instead of URLs.group_by/metrics/where plus optional query (measures/having/gap/ sequence/records as in sitetrax_query_assets) cover the span since the last send. send_time HH:MM in IANA timezone; days: daily, weekdays, weekend, weekly, or CSV. every_minutes (15-1440) replaces send_time and skips empty results. prompt: AI writes the email from the data. preview runs once, saves nothing. Omit/[] projects = all shared; expires_in_days=0 = 30 days. Mail goes only to the user's verified address. Requires the user's prior explicit confirmation.count means rows and distinct_assets means distinct stored Text unless eligibility establishes something narrower. group_by (maximum two dimensions) returns exact database groups; metrics or typed measures alone returns exact totals; neither returns paginated records with exact total_available. The separate totals block is exact for the resolved query window; displayed groups may be capped/truncated. Rows mode is record retrieval, not an analytics feed: never use one page or many, or code execution, to reconstruct analytics. Use structured where, aggregate having, conditional measures, bounded consecutive-event gap, server-side sequence, or a purpose tool. If none applies, report the capability gap. Flat text/status/type/camera filters combine with where; date inputs set the window; timezone controls buckets; fields/order/limit/offset control records. Omit projects or pass [] for all shared projects. fields is a database projection; limit/offset page records and never alter exact totals. gap is not general event pairing. sequence performs bounded ordered-stage duration/occupancy analysis for one eligible equipment family with explicit operational duration/pairing assumptions; USDOT is not one truck identity. Optional top-level records selects bounded native visits or event decisions with where/order/fields/paging and expected_result_version, without recalculating the aggregate or imposing a hardcoded top-N rule. Gate interpretation may be heading-based (asset_heading + l2risout) or project/camera-designated; use returned metadata and ask when ambiguity is material. Interpret type+status together and exclude sentinel/placeholder Text from equipment histories. Discover exact vocabulary and policies with sitetrax_schema(resource="query_capabilities"|"sequences"|"underlying_records"). Use sitetrax_reference_lookup for product/logistics definitions.group_by (maximum two dimensions) returns exact database groups and an exact totals block; metrics or mutually exclusive typed measures without grouping returns exact totals; neither returns paginated records with exact total_available. Displayed groups can be truncated and rows are retrieval only: row listings are paginated previews. Never exhaust them or use code execution to recreate an analytical result. Report a capability gap when no declared server reducer applies. Video dimensions include camera/camera_id, status, stream_type, project/project_id, and created_at day/week/month in timezone. Metrics are count, sum_assets, and avg_assets; typed measures add conditional counts and sum/avg/min/max over number_of_assets. Flat camera/status/stream_type filters combine with structured where; having filters requested aggregate measures. Date inputs bound created_at. order, limit, and offset affect record mode only. Omit projects or pass [] for all shared projects. Discover exact fields, operators, and limits with sitetrax_schema(resource="query_capabilities").status_codes replaces it and may include any code. Selection is not a genuine-concern classifier: interpret each row's emitted type and status together using sitetrax_schema(resource="asset_semantics"). Compact items omit type, so refetch the numeric detection ID or query it with type before classifying it. Compatibility keys total_scans_in_window, needs_review_count, and review_rate_percent mean stored rows and the selected-row share. They do not mean verified scans or verified concerns. Omit projects/pass [] for all shared projects; date inputs define the echoed resolved window; limit only caps listed candidates, not totals.Click any prompt to copy it.
Add this badge to your site or README. It links to this page.
HTML
<a href="https://mcp-app-store.keenethics-labs.com/app/sitetrax-io"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>
Markdown
[](https://mcp-app-store.keenethics-labs.com/app/sitetrax-io)