# AcuSight ## Docs - [What is AcuSight?](https://docs.acusight.io/what-is-acusight.md): Train computer-vision models and run them across a fleet of edge devices — from the first image on the line to a model in production. - [For data scientists](https://docs.acusight.io/get-started/personas/data-scientist.md): Annotate data, version datasets, train models, and deploy them — the model-building path through AcuSight. - [For infrastructure & ops](https://docs.acusight.io/get-started/personas/infra.md): Deploy AcuSight, provision edge devices, and run the fleet in production — the operations path. - [For developers & integrators](https://docs.acusight.io/get-started/personas/dev.md): Build on AcuSight's REST API and Co-Pilot/MCP. - [For explorers](https://docs.acusight.io/get-started/personas/explorer.md): Kick the tires and see what AcuSight does — the fastest path from zero to a running platform. - [Quickstart](https://docs.acusight.io/quickstart.md): Train and deploy your first model with AcuSight — the whole loop, end to end. - [Edge data collection](https://docs.acusight.io/guides/data/edge-data-collection.md): How AcuSight collects image data from edge devices in real time — and turns it into a project you can label. - [Projects & batches](https://docs.acusight.io/guides/data/projects-batches.md): Organize incoming images into projects and batches, ready to label and train. - [Annotation](https://docs.acusight.io/guides/data/annotation.md): Label images with bounding boxes and classes so a model can learn to detect them. - [Dataset versioning](https://docs.acusight.io/guides/data/dataset-versioning.md): Freeze your labeled images into an immutable, reproducible dataset version. - [Augmentation](https://docs.acusight.io/guides/data/augmentation.md): Expand and clean your training split with preprocessing and augmentation. - [Training & experiments](https://docs.acusight.io/guides/train-eval/training-experiments.md): Train a model on a dataset version and track its metrics live. - [Model eval & champion promotion](https://docs.acusight.io/guides/train-eval/model-eval-champion.md): Compare trained models and promote the best one to production. - [Connect a device](https://docs.acusight.io/guides/deploy/device-provisioning.md): Register an edge device with AcuSight from the dashboard and bring it online with a one-time setup command. - [Deploy a model to devices](https://docs.acusight.io/guides/deploy/deploy-model.md): Ship a trained model to the devices you choose and watch it run live. - [Fleet management](https://docs.acusight.io/guides/operate/fleet-management.md): Monitor and manage your edge devices — individually and in groups — from the AcuSight dashboard. - [Device settings](https://docs.acusight.io/guides/operate/device-settings.md): Configure an individual device from the dashboard — its video sources, analytics, detection zones, and network — and push the changes out remotely. - [Container management & logs](https://docs.acusight.io/guides/operate/container-management.md): Real container orchestration on every edge device — start, stop, inspect, stream logs, and open a console from the dashboard. - [Monitoring](https://docs.acusight.io/guides/operate/monitoring.md): Watch platform, fleet, and model health on the built-in dashboards. - [Recording & livestream](https://docs.acusight.io/guides/operate/recording-livestream.md): Record edge video, play it back on a timeline, and monitor devices live. - [Reliability](https://docs.acusight.io/guides/operate/reliability.md): What AcuSight runs for you as a hosted service, what stays yours, and how long data is kept. - [Orgs, users & multi-tenancy](https://docs.acusight.io/guides/administer/orgs-multitenancy.md): Administer your organization — how users join, roles, and managing members. - [Signing in](https://docs.acusight.io/guides/administer/signing-in.md): Create an AcuSight account with Google, GitHub, or email — and how your work email links you to your organization. - [Device keys & provisioning tokens](https://docs.acusight.io/guides/administer/device-keys.md): The two credentials in a device's life — the setup token that enrolls it and the device key it authenticates with — and how to rotate or revoke them. - [Security & trust](https://docs.acusight.io/guides/administer/security.md): How AcuSight protects your data — identities for people and devices, TLS everywhere, strict organization isolation, and short-lived media. - [REST API overview & auth](https://docs.acusight.io/guides/integrate/rest-api-overview.md): How to authenticate and call the AcuSight REST API. - [Co-Pilot & MCP](https://docs.acusight.io/guides/integrate/copilot-mcp.md): Use the built-in AcuSight Co-Pilot, and connect external AI tools to your AcuSight data over MCP. - [Architecture](https://docs.acusight.io/concepts/architecture.md): How the pieces of AcuSight fit together — a fleet of edge cameras running models on-device, one platform, and the dashboard, API, and Co-Pilot on top. - [Model lifecycle](https://docs.acusight.io/concepts/model-lifecycle.md): How a dataset becomes a trained model, gets promoted to champion, and ships to your fleet — with a clear path back. - [Multi-tenancy & org scoping](https://docs.acusight.io/concepts/multi-tenancy.md): How AcuSight keeps every organization's data isolated from every other's. - [Vision Language Models](https://docs.acusight.io/concepts/vlm.md): A demonstration of AcuSight's vision-language model capabilities — cameras that describe in plain English what they see. - [Co-Pilot architecture](https://docs.acusight.io/concepts/copilot.md): How the AcuSight Co-Pilot turns a plain-English question into an answer grounded in your live data. - [Config & environment variables](https://docs.acusight.io/reference/config-env.md): How AcuSight is configured, and where the configuration reference lives. - [CLI / make targets](https://docs.acusight.io/reference/cli-make-targets.md): Where the command-line tooling for running AcuSight lives. - [Glossary](https://docs.acusight.io/reference/glossary.md): The core nouns of AcuSight — projects, batches, datasets, models, and devices — defined in one place. - [Changelog](https://docs.acusight.io/reference/changelog.md): What's new in AcuSight. - [Troubleshooting](https://docs.acusight.io/reference/troubleshooting.md): Common issues and how to resolve them from the AcuSight dashboard. - [Get help](https://docs.acusight.io/reference/get-help.md): Where to go when you're stuck — support channels, what to include in a report, and how to follow known issues. - [Build a defect detector](https://docs.acusight.io/tutorials/defect-detector.md): A guided, worked example of the full AcuSight loop — from raw images to a model catching defects live on a device. - [Provision & manage an edge fleet](https://docs.acusight.io/tutorials/edge-fleet.md): Bring several edge devices online, organize them into groups, and deploy a model across all of them at once. - [Ask the Co-Pilot about your fleet](https://docs.acusight.io/tutorials/copilot.md): Use the built-in AcuSight Co-Pilot to answer questions about your devices, data, and models in plain English. - [API reference](https://docs.acusight.io/api-reference/api-introduction.md): Authenticate and explore the AcuSight REST API — generated from the platform's OpenAPI spec, always in sync with the running API. - [Agent — report deployment job phase progress](https://docs.acusight.io/api-reference/devices-&-provisioning/agent-—-report-deployment-job-phase-progress.md): Record a per-phase progress update from an edge agent. Upserts on (device_id, job_id) and appends to phase_history. - [List devices](https://docs.acusight.io/api-reference/devices-&-provisioning/list-devices.md): List all registered devices with optional filtering and pagination - [Register new device](https://docs.acusight.io/api-reference/devices-&-provisioning/register-new-device.md): Register a new edge device for management and monitoring - [Get device details](https://docs.acusight.io/api-reference/devices-&-provisioning/get-device-details.md): Retrieve device information by numeric ID or device identifier - [Delete device](https://docs.acusight.io/api-reference/devices-&-provisioning/delete-device.md): Delete a device and all associated data including batches, images, annotations, and Portainer endpoint - [Update device](https://docs.acusight.io/api-reference/devices-&-provisioning/update-device.md): Update device settings including name, auto-assignment, status, or Portainer configuration - [Discover device configuration](https://docs.acusight.io/api-reference/devices-&-provisioning/discover-device-configuration.md): Auto-discovers MediaMTX configuration via Portainer container inspection - [Set per-device config override](https://docs.acusight.io/api-reference/devices-&-provisioning/set-per-device-config-override.md): Stores a full-map per-device override for streamproc configs. Override key set MUST be a superset of the current release's rendered_configs keys. - [Discard per-device config override](https://docs.acusight.io/api-reference/devices-&-provisioning/discard-per-device-config-override.md) - [Set per-device model override](https://docs.acusight.io/api-reference/devices-&-provisioning/set-per-device-model-override.md): Stores a partial per-device override (role -> model_version_id). Only overridden roles appear; non-overridden roles follow the group's release. - [Discard per-device model override](https://docs.acusight.io/api-reference/devices-&-provisioning/discard-per-device-model-override.md) - [Get effective configs for a device](https://docs.acusight.io/api-reference/devices-&-provisioning/get-effective-configs-for-a-device.md) - [Get effective per-role models for a device](https://docs.acusight.io/api-reference/devices-&-provisioning/get-effective-per-role-models-for-a-device.md) - [Get the device's group and its assigned release](https://docs.acusight.io/api-reference/devices-&-provisioning/get-the-devices-group-and-its-assigned-release.md) - [List pipelines for a device](https://docs.acusight.io/api-reference/devices-&-provisioning/list-pipelines-for-a-device.md): List all pipelines belonging to a specific device - [Create pipeline](https://docs.acusight.io/api-reference/devices-&-provisioning/create-pipeline.md): Create a new video processing pipeline on an edge device - [Get pipeline config](https://docs.acusight.io/api-reference/devices-&-provisioning/get-pipeline-config.md): Returns the current config reported by the agent and the desired config set by the UI - [Set desired pipeline config](https://docs.acusight.io/api-reference/devices-&-provisioning/set-desired-pipeline-config.md): Sets the desired config that will be pushed to the edge device on next checkin - [MSE playback init](https://docs.acusight.io/api-reference/devices-&-provisioning/mse-playback-init.md) - [Recording playback window manifest](https://docs.acusight.io/api-reference/devices-&-provisioning/recording-playback-window-manifest.md) - [List playback spans](https://docs.acusight.io/api-reference/devices-&-provisioning/list-playback-spans.md) - [Span segment bytes (fmp4)](https://docs.acusight.io/api-reference/devices-&-provisioning/span-segment-bytes-fmp4.md) - [Get per-device streamproc state](https://docs.acusight.io/api-reference/devices-&-provisioning/get-per-device-streamproc-state.md) - [Update device status](https://docs.acusight.io/api-reference/devices-&-provisioning/update-device-status.md): Update device approval status (approved or rejected) - [Forward browser WHEP offer to device](https://docs.acusight.io/api-reference/devices-&-provisioning/forward-browser-whep-offer-to-device.md): Forwards the SDP offer to the device via its signalling WebSocket, returns the agent's SDP answer on success. - [Get pipeline](https://docs.acusight.io/api-reference/devices-&-provisioning/get-pipeline.md): Get details of a specific pipeline - [Delete pipeline](https://docs.acusight.io/api-reference/devices-&-provisioning/delete-pipeline.md): Delete a pipeline and cascade to associated batches, images, and actions - [Update pipeline](https://docs.acusight.io/api-reference/devices-&-provisioning/update-pipeline.md): Update pipeline configuration and settings - [Get recording-retention settings](https://docs.acusight.io/api-reference/devices-&-provisioning/get-recording-retention-settings.md) - [Update recording-retention settings](https://docs.acusight.io/api-reference/devices-&-provisioning/update-recording-retention-settings.md) - [Device health check-in](https://docs.acusight.io/api-reference/devices-&-provisioning/device-health-check-in.md): Reports device health status and model state. Returns deployment instructions if desired state differs from actual state. - [Upload image from edge device](https://docs.acusight.io/api-reference/devices-&-provisioning/upload-image-from-edge-device.md): Receives an image file and optional metadata from a cloud-mode edge device, - [Forward device data to pipeline-data ingest](https://docs.acusight.io/api-reference/devices-&-provisioning/forward-device-data-to-pipeline-data-ingest.md): Proxies POST /api/device/ingest to POST http://pipeline-data:8082/api/pipeline-data/ingest, - [Get model metadata for device deployment](https://docs.acusight.io/api-reference/devices-&-provisioning/get-model-metadata-for-device-deployment.md): Returns model metadata including name, version, file extension, task type, and label map. - [Download model artifact](https://docs.acusight.io/api-reference/devices-&-provisioning/download-model-artifact.md): Streams the model artifact file from MinIO to the requesting device. - [Upload recording segment (fmp4)](https://docs.acusight.io/api-reference/devices-&-provisioning/upload-recording-segment-fmp4.md) - [Get device self info](https://docs.acusight.io/api-reference/devices-&-provisioning/get-device-self-info.md): Returns information about the authenticated device. Used by edge agents to verify their registration and get organization context. - [Device WebSocket signalling channel](https://docs.acusight.io/api-reference/devices-&-provisioning/device-websocket-signalling-channel.md): Long-lived WebSocket used to forward WHEP SDP offers from browsers to the device's local MediaMTX. Device authenticates via X-Device-Key. - [WebSocket proxy to pipeline-data stream ingest](https://docs.acusight.io/api-reference/devices-&-provisioning/websocket-proxy-to-pipeline-data-stream-ingest.md): Upgrades GET /api/device/stream to a WebSocket and proxies bidirectionally - [Provision a new device](https://docs.acusight.io/api-reference/devices-&-provisioning/provision-a-new-device.md): Provisions a new edge device using a valid provisioning token. Creates the device record, generates an API key, and optionally creates a Portainer Edge endpoint. This endpoint is called by the setup script on edge devices. - [Get current authenticated member](https://docs.acusight.io/api-reference/devices-&-provisioning/get-current-authenticated-member.md) - [List organization members](https://docs.acusight.io/api-reference/devices-&-provisioning/list-organization-members.md) - [Remove organization member](https://docs.acusight.io/api-reference/devices-&-provisioning/remove-organization-member.md) - [Update member role](https://docs.acusight.io/api-reference/devices-&-provisioning/update-member-role.md) - [List ML coordinator credentials](https://docs.acusight.io/api-reference/devices-&-provisioning/list-ml-coordinator-credentials.md): Lists the long-lived coordinator credentials for the caller's organization (plus all-org credentials for platform admins). The raw credential is never returned. - [Revoke an ML coordinator credential](https://docs.acusight.io/api-reference/devices-&-provisioning/revoke-an-ml-coordinator-credential.md): Revokes a long-lived coordinator credential by id. Org admins may revoke their organization's credentials; all-org credentials require a platform admin. - [Serve the ML appliance installer script](https://docs.acusight.io/api-reference/devices-&-provisioning/serve-the-ml-appliance-installer-script.md): Returns a shell script that can be piped to sh to install an AcuSight ML appliance. Usage: `curl -sSL /ml-setup | sh -s -- ami_` - [Serve the ML appliance GPU compose overlay](https://docs.acusight.io/api-reference/devices-&-provisioning/serve-the-ml-appliance-gpu-compose-overlay.md) - [Serve the ML appliance docker-compose.yml](https://docs.acusight.io/api-reference/devices-&-provisioning/serve-the-ml-appliance-docker-composeyml.md) - [Serve the ML appliance env template](https://docs.acusight.io/api-reference/devices-&-provisioning/serve-the-ml-appliance-env-template.md) - [List provisioning tokens](https://docs.acusight.io/api-reference/devices-&-provisioning/list-provisioning-tokens.md): Returns all provisioning tokens for the current organization, sorted by creation date (newest first) - [Create a provisioning token](https://docs.acusight.io/api-reference/devices-&-provisioning/create-a-provisioning-token.md): Creates a new provisioning token that can be used to onboard edge devices. The raw token is only returned once and should be saved securely. - [Revoke a provisioning token](https://docs.acusight.io/api-reference/devices-&-provisioning/revoke-a-provisioning-token.md): Revokes a provisioning token, preventing it from being used to provision new devices - [List active alerts](https://docs.acusight.io/api-reference/alerts-&-telemetry/list-active-alerts.md): Retrieves active model monitoring alerts from Prometheus with optional model filtering. - [List alert rules](https://docs.acusight.io/api-reference/alerts-&-telemetry/list-alert-rules.md): Lists custom alert rules configured for model monitoring. - [Create alert rule](https://docs.acusight.io/api-reference/alerts-&-telemetry/create-alert-rule.md): Creates a Prometheus alert rule used for model monitoring. - [Delete alert rule](https://docs.acusight.io/api-reference/alerts-&-telemetry/delete-alert-rule.md): Deletes a custom alert rule and reloads Prometheus configuration. - [Create ticket for alert](https://docs.acusight.io/api-reference/alerts-&-telemetry/create-ticket-for-alert.md): Creates a support ticket for an alert - [Process Alertmanager webhook](https://docs.acusight.io/api-reference/alerts-&-telemetry/process-alertmanager-webhook.md): Receives and processes alerts from Alertmanager - [Process RUM event](https://docs.acusight.io/api-reference/alerts-&-telemetry/process-rum-event.md): Receives and processes Real User Monitoring events from the frontend - [Process batch RUM events](https://docs.acusight.io/api-reference/alerts-&-telemetry/process-batch-rum-events.md): Receives and processes multiple Real User Monitoring events from the frontend - [Update virtual device configuration](https://docs.acusight.io/api-reference/virtual-devices/update-virtual-device-configuration.md): Update simulation parameters for a virtual device (must be stopped) - [Pause virtual device simulation](https://docs.acusight.io/api-reference/virtual-devices/pause-virtual-device-simulation.md): Temporarily pause data generation simulation - [Start virtual device simulation](https://docs.acusight.io/api-reference/virtual-devices/start-virtual-device-simulation.md): Start data generation simulation for an approved virtual device - [Get virtual device statistics](https://docs.acusight.io/api-reference/virtual-devices/get-virtual-device-statistics.md): Get current simulation statistics including images uploaded and progress - [Stop virtual device simulation](https://docs.acusight.io/api-reference/virtual-devices/stop-virtual-device-simulation.md): Stop data generation simulation for a virtual device - [Create, approve, and start virtual device(s)](https://docs.acusight.io/api-reference/virtual-devices/create-approve-and-start-virtual-devices.md): Create one or more virtual devices, automatically approve them, and start simulation in a single API call. Supports three modes: event (synthetic events), image (dataset uploads), or hybrid (both). For image mode, use either dataset_version_id (imported dataset) or dataset_path (filesystem). Use man… - [Get virtual device event type configurations](https://docs.acusight.io/api-reference/virtual-devices/get-virtual-device-event-type-configurations.md): Get all available event type presets for virtual devices, grouped by category (Manufacturing, Generic, Custom) - [Notify edge image upload](https://docs.acusight.io/api-reference/datasets-&-data/notify-edge-image-upload.md): Records metadata for an image uploaded by an edge device, updates or creates the working batch, and persists inline annotations when provided. - [Get annotation summary](https://docs.acusight.io/api-reference/datasets-&-data/get-annotation-summary.md): Returns aggregate annotation counts grouped by class and model. - [Delete annotation](https://docs.acusight.io/api-reference/datasets-&-data/delete-annotation.md): Deletes the annotation and updates the image annotation status if no annotations remain. - [Update annotation](https://docs.acusight.io/api-reference/datasets-&-data/update-annotation.md): Updates class value, coordinates, confidence, geometry, or z-order for an annotation. - [List batches](https://docs.acusight.io/api-reference/datasets-&-data/list-batches.md): Lists batches across projects with optional filtering by project, stage, source, device, status, and supports pagination and sorting. - [Get batch](https://docs.acusight.io/api-reference/datasets-&-data/get-batch.md): Retrieves batch metadata including project assignment and status. - [Delete batch](https://docs.acusight.io/api-reference/datasets-&-data/delete-batch.md): Queues an asynchronous job to delete the batch, related images, annotations, and storage objects. - [Update batch](https://docs.acusight.io/api-reference/datasets-&-data/update-batch.md): Central endpoint for updating batch lifecycle fields such as stage transitions, project assignment, assignees, and metadata. - [Delete batch annotations](https://docs.acusight.io/api-reference/datasets-&-data/delete-batch-annotations.md): Enqueues an asynchronous job to delete all annotations for the batch. - [Extract reviewed images](https://docs.acusight.io/api-reference/datasets-&-data/extract-reviewed-images.md): Extracts all images with annotation_status='completed' from the source batch into a new batch. The new batch inherits the project assignment and is in the 'annotating' stage. - [Notify batch images](https://docs.acusight.io/api-reference/datasets-&-data/notify-batch-images.md): Registers images that were uploaded out-of-band (e.g., CLI or tooling) and optionally ingests YOLO annotation files and class maps. - [Split batch images](https://docs.acusight.io/api-reference/datasets-&-data/split-batch-images.md): Assigns train/val/test splits for images in the batch while respecting manual assignments. By default, batches are promoted to the dataset stage unless `promote_to_dataset` is set to false. - [Get batch split summary](https://docs.acusight.io/api-reference/datasets-&-data/get-batch-split-summary.md): Returns existing split counts, lock counts, and per-class distribution metadata for a batch to assist UI split planning. Optionally filter by annotation_status. - [Upload batch file](https://docs.acusight.io/api-reference/datasets-&-data/upload-batch-file.md): Streams a file into object storage under the batch upload prefix. Supports optional path overrides for directory nesting. - [Stream batch files](https://docs.acusight.io/api-reference/datasets-&-data/stream-batch-files.md): Upload multiple files in a single multipart request with real-time progress via WebSocket - [Stream version image data](https://docs.acusight.io/api-reference/datasets-&-data/stream-version-image-data.md): Streams the stored version image bytes directly to the caller from the version MinIO prefix. - [Get dataset version](https://docs.acusight.io/api-reference/datasets-&-data/get-dataset-version.md): Retrieves dataset version details including batches, classes, preview images, and split statistics. - [Delete dataset version](https://docs.acusight.io/api-reference/datasets-&-data/delete-dataset-version.md): Deletes the dataset version metadata and underlying export artifacts. - [Export calibration dataset artifact](https://docs.acusight.io/api-reference/datasets-&-data/export-calibration-dataset-artifact.md): Enqueues an asynchronous job to export a capped YOLO dataset artifact for model calibration. - [Export dataset version](https://docs.acusight.io/api-reference/datasets-&-data/export-dataset-version.md): Enqueues an asynchronous job to export a dataset version in the requested format. - [Download dataset export](https://docs.acusight.io/api-reference/datasets-&-data/download-dataset-export.md): Streams the archive generated by an export job once it has completed. - [List drift alerts](https://docs.acusight.io/api-reference/datasets-&-data/list-drift-alerts.md): Lists all drift alerts with optional filtering by device, model, status, severity, and time range - [Get drift alert statistics](https://docs.acusight.io/api-reference/datasets-&-data/get-drift-alert-statistics.md): Returns aggregate statistics for drift alerts (for dashboard KPIs) - [Get drift alert](https://docs.acusight.io/api-reference/datasets-&-data/get-drift-alert.md): Returns a single drift alert with all details - [Update drift alert status](https://docs.acusight.io/api-reference/datasets-&-data/update-drift-alert-status.md): Updates the status of a drift alert to acknowledged or resolved - [List drift configurations](https://docs.acusight.io/api-reference/datasets-&-data/list-drift-configurations.md): Lists all drift configurations with optional filtering by level - [Create drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/create-drift-configuration.md): Creates a new drift configuration for a model, fleet, or device - [Get system default drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/get-system-default-drift-configuration.md): Returns the system default values for drift detection thresholds - [Get effective drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/get-effective-drift-configuration.md): Returns the fully resolved drift configuration for a device, applying hierarchy inheritance - [Get drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/get-drift-configuration.md): Returns a single drift configuration by ID - [Update drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/update-drift-configuration.md): Updates an existing drift configuration - [Delete drift configuration](https://docs.acusight.io/api-reference/datasets-&-data/delete-drift-configuration.md): Deletes a drift configuration - [Health check](https://docs.acusight.io/api-reference/datasets-&-data/health-check.md): Returns health status of the Data service - [List images](https://docs.acusight.io/api-reference/datasets-&-data/list-images.md): Returns images with optional filters for project, batch, device, annotation status, dataset split, class values, and text search. Supports eager annotation loading and inline streaming URLs. - [Get image](https://docs.acusight.io/api-reference/datasets-&-data/get-image.md): Returns image details, associated batch/project info, and optionally a streaming URL. - [Delete image](https://docs.acusight.io/api-reference/datasets-&-data/delete-image.md): Deletes an image record, all associated annotations, and removes the file from storage. - [Update image](https://docs.acusight.io/api-reference/datasets-&-data/update-image.md): Updates image attributes such as annotation status or dataset split assignments. - [List image annotations](https://docs.acusight.io/api-reference/datasets-&-data/list-image-annotations.md): Returns annotations ordered by z-order for the specified image. - [Create annotation](https://docs.acusight.io/api-reference/datasets-&-data/create-annotation.md): Creates an annotation for the specified image using normalized coordinates or geometry data. - [Reorder image annotations](https://docs.acusight.io/api-reference/datasets-&-data/reorder-image-annotations.md): Applies a new ordering for annotations on an image by updating their z-order positions. - [Stream image data](https://docs.acusight.io/api-reference/datasets-&-data/stream-image-data.md): Streams the stored image bytes directly to the caller. - [Get image URL](https://docs.acusight.io/api-reference/datasets-&-data/get-image-url.md): Generates an API URL clients can use to stream the image contents. - [Get job status](https://docs.acusight.io/api-reference/datasets-&-data/get-job-status.md): Retrieves the current status, message, and optionally progress/result details for an async job. - [List projects](https://docs.acusight.io/api-reference/datasets-&-data/list-projects.md): Returns paginated projects ordered by creation date. Provides lightweight DTOs suitable for list views. - [Create project](https://docs.acusight.io/api-reference/datasets-&-data/create-project.md): Creates a project with the specified labeling type and optional description. - [Get project](https://docs.acusight.io/api-reference/datasets-&-data/get-project.md): Returns detailed project information including associated batches, dataset versions, and classes. - [Delete project](https://docs.acusight.io/api-reference/datasets-&-data/delete-project.md): Deletes the project along with associated batches, versions, and classes. - [Update project](https://docs.acusight.io/api-reference/datasets-&-data/update-project.md): Updates name or description for a project. Project type is immutable and cannot be changed after creation. - [Get project analytics](https://docs.acusight.io/api-reference/datasets-&-data/get-project-analytics.md): Returns the latest dataset analytics for a project and can queue a regeneration job when requested. - [Get analytics dimensions](https://docs.acusight.io/api-reference/datasets-&-data/get-analytics-dimensions.md): Provides image dimension metrics along with sampled class labels for visualization. - [Get analytics histogram](https://docs.acusight.io/api-reference/datasets-&-data/get-analytics-histogram.md): Provides annotated image count distribution buckets for visualization. - [Create project batch](https://docs.acusight.io/api-reference/datasets-&-data/create-project-batch.md): Creates an upload batch scoped to a project, defaulting the name when not provided. - [List project classes](https://docs.acusight.io/api-reference/datasets-&-data/list-project-classes.md): Returns all classes associated with the project along with how many annotations use each class. - [Create project class](https://docs.acusight.io/api-reference/datasets-&-data/create-project-class.md): Adds a class to the project with an optional color. Class names must be unique per project. - [Update project class](https://docs.acusight.io/api-reference/datasets-&-data/update-project-class.md): Updates class name or color. When delete_class is true, removes the class and associated annotations for the project. - [List dataset versions](https://docs.acusight.io/api-reference/datasets-&-data/list-dataset-versions.md): Returns paginated dataset versions with image, batch, and class counts. - [Create dataset version](https://docs.acusight.io/api-reference/datasets-&-data/create-dataset-version.md): Validates the working dataset and enqueues an asynchronous job to create a new dataset version. - [Add batches to working dataset](https://docs.acusight.io/api-reference/datasets-&-data/add-batches-to-working-dataset.md): Validates batches belong to the project and moves them into the dataset stage. - [Create bulk annotations from inference](https://docs.acusight.io/api-reference/datasets-&-data/create-bulk-annotations-from-inference.md): Creates annotations from model predictions with auto-class creation - [Get project stats](https://docs.acusight.io/api-reference/datasets-&-data/get-project-stats.md): Provides counts of batches and images across workflow stages for a project. - [Subscribe to feed updates](https://docs.acusight.io/api-reference/datasets-&-data/subscribe-to-feed-updates.md): Establishes a WebSocket connection that streams live dataset and job updates to the client. - [Prepare calibration dataset](https://docs.acusight.io/api-reference/models-&-training/prepare-calibration-dataset.md): Exports a dataset version for model calibration (used by worker) - [Update export status](https://docs.acusight.io/api-reference/models-&-training/update-export-status.md): Updates export job status and artifact metadata - [Get inference request](https://docs.acusight.io/api-reference/models-&-training/get-inference-request.md): Returns inference results, thresholds, and completion metadata for an inference request. - [Update inference results](https://docs.acusight.io/api-reference/models-&-training/update-inference-results.md): Receives prediction results from the worker and processes them (e.g., creating annotations for auto-annotate jobs). - [Get job metrics](https://docs.acusight.io/api-reference/models-&-training/get-job-metrics.md): Returns metric history for a training job. Optionally filter by comma-separated keys. - [Batch insert metrics](https://docs.acusight.io/api-reference/models-&-training/batch-insert-metrics.md): Inserts multiple metric data points for a training job. - [Get latest metrics for job](https://docs.acusight.io/api-reference/models-&-training/get-latest-metrics-for-job.md): Returns only the most recent value for each metric key recorded for a job. - [Get model](https://docs.acusight.io/api-reference/models-&-training/get-model.md): Returns model version metadata including aliases, metrics and deployment counts. - [Download model weights](https://docs.acusight.io/api-reference/models-&-training/download-model-weights.md): Generates a short-lived URL for downloading model weights from object storage. - [Export model](https://docs.acusight.io/api-reference/models-&-training/export-model.md): Creates an export job for a model version and returns the job record. - [List monitored devices](https://docs.acusight.io/api-reference/models-&-training/list-monitored-devices.md): Proxies to the core service to list devices currently marked as deployed. - [List monitored models](https://docs.acusight.io/api-reference/models-&-training/list-monitored-models.md): Aggregates deployment and inference metrics for models over the requested time range. - [Auto-annotate images](https://docs.acusight.io/api-reference/models-&-training/auto-annotate-images.md): Creates an auto-annotation inference request that applies model predictions to images. - [Create inference request](https://docs.acusight.io/api-reference/models-&-training/create-inference-request.md): Creates an inference request for the given project and queues it for asynchronous processing. - [List models](https://docs.acusight.io/api-reference/models-&-training/list-models.md): Returns paginated model versions for a project with optional alias filtering. - [Get Project Settings](https://docs.acusight.io/api-reference/models-&-training/get-project-settings.md): Returns ML-related settings for a project including auto-deploy configuration. - [Update Project Settings](https://docs.acusight.io/api-reference/models-&-training/update-project-settings.md): Updates ML-related settings for a project including auto-deploy configuration. - [Create auto training job](https://docs.acusight.io/api-reference/models-&-training/create-auto-training-job.md): Creates a snapshot with default augmentations and triggers a training job. - [Create instant training job](https://docs.acusight.io/api-reference/models-&-training/create-instant-training-job.md): Creates a snapshot and triggers a quick training job on small datasets. - [Get training logs](https://docs.acusight.io/api-reference/models-&-training/get-training-logs.md): Returns aggregated logs recorded for a training job. - [Get organization details](https://docs.acusight.io/api-reference/organization/get-organization-details.md): Returns details about the current user's organization including device and member counts - [Create organization](https://docs.acusight.io/api-reference/organization/create-organization.md) - [Update organization](https://docs.acusight.io/api-reference/organization/update-organization.md): Updates the organization name - [List default-group rules](https://docs.acusight.io/api-reference/organization/list-default-group-rules.md): Returns all per-org default-group rules for the caller's organization - [Set/upsert a default-group rule](https://docs.acusight.io/api-reference/organization/setupsert-a-default-group-rule.md): Sets the edge group that newly provisioned matching devices are auto-added to - [Clear a default-group rule](https://docs.acusight.io/api-reference/organization/clear-a-default-group-rule.md): Deletes the default-group rule for the given scope (defaults to the org-wide catch-all) - [Correlate quality events with process parameters](https://docs.acusight.io/api-reference/pipeline-data/correlate-quality-events-with-process-parameters.md): Returns per-quality_event tuples joined to the process_params metric that produced them, with a cycle_counter|time_window join_method. - [Aggregate detection statistics](https://docs.acusight.io/api-reference/pipeline-data/aggregate-detection-statistics.md): Returns per-class detection counts, average confidence, and stddev for drift detection. - [Get bounding box statistics](https://docs.acusight.io/api-reference/pipeline-data/get-bounding-box-statistics.md): Returns per-class bounding box size and aspect ratio statistics for drift detection. - [Get per-frame detection statistics](https://docs.acusight.io/api-reference/pipeline-data/get-per-frame-detection-statistics.md): Returns statistics about detection counts per frame for drift detection. - [Clear/acknowledge event rows](https://docs.acusight.io/api-reference/pipeline-data/clearacknowledge-event-rows.md): Marks one or more event rows (by type + id) as cleared so they are excluded from queries. Single ({type,id}) or bulk ({type,ids[]}, max 500). - [Get map config for a device](https://docs.acusight.io/api-reference/pipeline-data/get-map-config-for-a-device.md): Returns the floor plan configuration (zones, cameras) for the specified device. - [Upsert map config for a device](https://docs.acusight.io/api-reference/pipeline-data/upsert-map-config-for-a-device.md): Creates or updates the floor plan configuration (zones, cameras) for a device. - [Get model performance metrics](https://docs.acusight.io/api-reference/pipeline-data/get-model-performance-metrics.md): Returns per-model statistics including detection count, confidence stats, unique labels, and device count. - [Aggregate quality event counts](https://docs.acusight.io/api-reference/pipeline-data/aggregate-quality-event-counts.md): Returns per-(source, disposition, defect_code[, part_number]) event counts for shadow-mode A/B comparison. - [Query windowed pipeline data](https://docs.acusight.io/api-reference/pipeline-data/query-windowed-pipeline-data.md): Returns events or metrics within the requested time window, with optional id cursor and filters. - [Stream pipeline data forward-only](https://docs.acusight.io/api-reference/pipeline-data/stream-pipeline-data-forward-only.md): Streams events or metrics after a cursor (ingestion_time|id) with optional end boundary. - [Supported edge devices](https://docs.acusight.io/hardware/supported-devices.md): What it takes to run the AcuSight edge agent, and the hardware platforms it's validated on. - [Board bring-up](https://docs.acusight.io/hardware/board-bringup.md): Prerequisites for accelerated boards — get the vendor's base image booted and online before installing the AcuSight edge agent. - [Edge agent](https://docs.acusight.io/hardware/edge-agent.md): The lightweight containerized agent that runs on every AcuSight edge device — what it does and how it connects. - [Air-gapped / on-prem](https://docs.acusight.io/hardware/air-gapped.md): Running AcuSight in air-gapped and on-prem environments. - [Project setups](https://docs.acusight.io/project/overview.md): A working space to document AcuSight deployment and project setups — configurations, recipes, and site-specific notes — as they're established. - [Setup template](https://docs.acusight.io/project/setups/template.md): Copy this template to document a new project or deployment setup. ## OpenAPI Specs - [api-spec](https://docs.acusight.io/api-spec.json) - [openapi](https://docs.acusight.io/api-reference/openapi.json) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.