# Roboflow MCP connector

The Roboflow connector brings your computer vision workspaces, projects, dataset versions, training jobs, and images to Claude, ChatGPT, Cursor, and any MCP client, so your team can ask about them and act on them in plain language.

Source: https://elaichi.ai/connectors/roboflow/

## Facts

| | |
| --- | --- |
| Application | Roboflow |
| Category | Artificial Intelligence |
| AI tools | 103 |
| Authentication | Connects with an API key |
| Needs your own OAuth app | No |
| MCP endpoint | https://api.elaichi.ai/mcp |
| Works with | Claude, ChatGPT, Cursor, any MCP client, and the Elaichi Agent |
| Tools advertised by name | No. 103 tools is past the 30-tool threshold, so clients use `search_tools` and `execute_tool` |

## What you can ask once Roboflow is connected

- Which Roboflow projects added the most images last week?
- List versions ready to export in the detection workspace.
- Summarize project health across all Roboflow workspaces.

## Connect Roboflow in Elaichi

This happens once for the organization, before any client is involved.

1. Open Connections, choose Add connection, and pick Roboflow.
2. Optionally set Share with, then press Connect.
3. Paste a Roboflow API key. One person generates a token in Roboflow and pastes it once. Everyone else works through Share with, and never sees it.

Credentials are vaulted and nobody, including the AI, reads them back. The connection becomes a toolbox immediately, so you can curate which Roboflow tools are exposed, rename them, or freeze arguments before anyone points a client at it.

## Connect Roboflow to Claude

Endpoint: https://api.elaichi.ai/mcp

1. Open Customize, then Connectors.
2. Press Add.
3. Name it, paste the MCP server URL, then Continue.
4. Sign in and approve.

On Team and Enterprise, an Owner adds it once. Everyone else turns it on for themselves.

## Connect Roboflow to ChatGPT

Endpoint: https://api.elaichi.ai/mcp

1. Open Plugins, then press the + button.
2. Name it and paste the endpoint into Server URL.
3. Leave Authentication on OAuth, then tick the risk acknowledgement.
4. Press Create, then sign in and approve.

Works on the web today. The plugin directory lives at chatgpt.com/plugins.

## Connect Roboflow to Cursor

Endpoint: https://api.elaichi.ai/mcp

1. Open `~/.cursor/mcp.json`.
2. Add the endpoint under `mcpServers`.
3. Reload Cursor, then sign in and approve.

Set up per machine, so repeat it on each computer you work from.

## Connect Roboflow to any MCP client

Endpoint: https://api.elaichi.ai/mcp

1. Add the endpoint as a remote MCP server.
2. Sign in and approve.

The Elaichi Agent already has these tools, with nothing to set up.

## What the consent screen decides

Only Read is granted by default, which is not enough to call a Roboflow tool. Over MCP there is no trusted place to confirm a write in the moment, so the consent screen is the standing approval rather than a formality. Grant Read and Run tools. Think hard before granting Delete, which reaches into connected apps and cannot be undone.

## What teams do with Roboflow through Elaichi

### Kick off a training run without opening a tab

Machine learning. Ask for a training job on the latest version of a project and check on it later, or stop it if the early numbers look wrong.

### Upload new images into the right project

Data annotation. Send a batch of images straight into the Roboflow project they belong to, then confirm they landed by asking for them back by ID.

### Compare dataset versions before you ship

Computer vision engineering. Pull up every version of a project, see which ones have exports and trained models, and pick the one to promote.

### Check project health across the workspace

Operations. Get a plain read on which projects in a Roboflow workspace are in good shape and which are missing images, labels, or a trained model.

### Fork a project to try a new idea

Product. Copy an existing project into a fresh fork so a new detection experiment starts from real labeled data instead of an empty dataset.

### Keep the workspace tidy

Research. List every project and version in a workspace, spot the abandoned experiments, and remove the ones nobody trained in months.

## Frequently asked questions

### How do I connect Roboflow to Claude?

Connect Roboflow in Elaichi first by pasting the API key from your Roboflow account settings, there is no OAuth application to register and no client ID or secret to generate. Then in Claude open Customize, then Connectors, then Add, and paste https://api.elaichi.ai/mcp as the endpoint. Sign in to Elaichi when Claude asks, and your Roboflow workspaces, projects, and training jobs are available in the conversation.

### Does Roboflow work with ChatGPT and Cursor as well as Claude?

Yes. Roboflow is connected once in Elaichi, and the same endpoint, https://api.elaichi.ai/mcp, works in Claude, ChatGPT, Cursor, any other MCP client, and the Elaichi Agent. You do not connect Roboflow separately for each client.

### What can an AI agent actually do with my Roboflow data?

With Roboflow connected, an agent can list your workspaces and projects, open a specific project or version, check project health, see which versions have exports and trained models, and upload images into a project. It can also create projects, fork an existing one, start a training job, and cancel or stop a run that is going badly. Because Roboflow exposes many actions, short concrete asks like "start training on version 4 of the pallet project" work better than long paragraphs.

### Does connecting Roboflow give the AI access to every workspace?

No. The AI acts inside the access of the Roboflow API key that was connected and the person who signed in, so it sees only the workspaces, projects, and versions that person could already open in Roboflow. Elaichi can narrow that further, for example to reading projects but never deleting them, and it never widens what the person could already do.

### Can my team share one Roboflow connection?

Yes. One person connects Roboflow with their API key and shares the connection with a team in Elaichi, and nobody else ever sees or handles the key. Each teammate still signs in to Elaichi as themselves, so the audit log records exactly who uploaded an image, started a training job, or deleted a version.

### Can I stop an agent from deleting or changing things in Roboflow?

Yes. Restrictions in Elaichi apply per action, so you can allow reading Roboflow projects and versions while blocking deletion of projects, deletion of versions, or starting training jobs. A blocked action is never shown to Claude, ChatGPT, or Cursor at all, so no prompt, accidental or otherwise, can reach it.

### What happens to a Roboflow connection when someone leaves?

Offboarding a person in Elaichi ends their access to Roboflow through every client at once. A Roboflow connection they shared with the team keeps working for everyone else, because the credential lives in Elaichi and not on their laptop. If you want to remove Roboflow entirely, disconnecting it once in Elaichi removes it from Claude, ChatGPT, Cursor, and every other client together.

## All 103 Roboflow tools

Every tool below is callable through https://api.elaichi.ai/mcp once Roboflow is connected, subject to the toolbox it is in and the restrictions on the caller.

- **List all Roboflow workspace** (List). Verify your Roboflow API key and retrieve the authenticated workspace identifier from the root endpoint. Returns: welcome, instructions, docs, workspace.
- **Get single Roboflow workspace by ID** (Get). Get a Roboflow workspace by id, returning its metadata and full project list. Returns: name, url, members, projects, config, patch. Required: id.
- **Create a Roboflow workspace search** (Create). Search images across a Roboflow workspace using a query string that supports filtering, sorting, and semantic search. Returns: workflows, status. Required: workspace, query.
- **Create a Roboflow project** (Create). Create a new project in a Roboflow workspace. Returns: id, name, created, updated, images, unannotated, annotation, versions, public, splits, colors, classes, icon. Required: workspace, name.
- **List all Roboflow project** (List). List all projects in a Roboflow workspace. Returns each project's id, type, name, annotation, created, updated, images, unannotated, versions, public, splits, and classes. Required: workspace.
- **Delete a Roboflow project by ID** (Delete). Move a Roboflow project to Trash (soft delete; items are retained for 30 days before permanent cleanup). Returns: deleted, workspace, project, projectId, trash. Required: workspace, id.
- **Get single Roboflow project by ID** (Get). Get a Roboflow project and its dataset versions by id. Returns the project object including id, name, type, annotation, created, updated, images, unannotated, public, splits, classes, and a versions array. Required: workspace, id.
- **Create a Roboflow project fork** (Create). Fork a public Roboflow Universe project into a target workspace. The operation runs asynchronously and returns a taskId and polling url you can use to check fork status. Returns: taskId, url. Required: workspace, plus either url (full Universe project URL) or source_project (project slug) — at least one must be supplied.
- **List all Roboflow project health** (List). Get the dataset health check for a Roboflow project. Returns a summary of annotation quality and image statistics including health, stats.numImages, stats.numAnnotations, stats.missingAnnotations, stats.nullAnnotations, and stats.classDistribution. Required: workspace, project.
- **List all Roboflow project model** (List). List all trained models in a Roboflow project, including both version-trained models and standalone NAS-generated models. Returns: url, version, name, modelType, created, train, metrics, nasFamily, group, favorites, recommended. Required: workspace, project. Optionally filter to a single NAS run by passing group.
- **List all Roboflow version** (List). Get a specific Roboflow dataset version by workspace, project, and version number. Returns the full version record including workspace (name, url, members), project (id, name, type, images), and version (id, name, created, images, model, preprocessing, augmentation, exports, train). Required: workspace, project, version.
- **Delete a Roboflow version by ID** (Delete). Move a Roboflow dataset version to Trash by id. Returns: deleted, workspace, project, projectId, version, trash. Items are retained in Trash for 30 days before automatic cleanup; permanent deletion is available only via the Roboflow web app. Required: workspace, project, id.
- **Get single Roboflow version by ID** (Get). Get a specific Roboflow dataset version by id. Returns the full version record including workspace (name, url, members), project (id, name, type, images), and version (id, name, created, images, model, preprocessing, augmentation, exports, train). Required: workspace, project, id.
- **List all Roboflow version export** (List). Download a Roboflow dataset version in a specific export format. Returns workspace, project, and version details including version id, images count, and the list of generated export formats. Required: workspace, project, version, format.
- **Create a Roboflow training job** (Create). Create a Roboflow training job for a dataset version. Returns: jobId, status, version, createdAt. Training runs asynchronously; poll the job endpoint to check its status. Required: workspace, project, version.
- **List all Roboflow training job** (List). List Roboflow training jobs for a dataset version. Returns each job's id, status, version, model_type, createdAt, and progress. Required: workspace, project, version.
- **Get single Roboflow training job by ID** (Get). Get a single Roboflow training job by id. Returns the job's id, status, version, model_type, createdAt, and progress. Status values: queued, training, complete, failed, cancelled. Required: workspace, project, version, id.
- **Roboflow training job cancel** (Cancel). Cancel a running Roboflow training job for a dataset version. Returns: success, jobId, status, version, createdAt. Required: workspace, project, version.
- **Roboflow training job stop** (Stop). Stop a running Roboflow training job for a dataset version. Returns: success, jobId, status, version, createdAt. Required: workspace, project, version.
- **Create a Roboflow version training** (Create). Start a training job on a Roboflow dataset version. Training runs asynchronously; the response confirms the job was queued, not that it has completed. Returns: jobId, status, version, createdAt, success. Required: workspace, project, version.
- **Create a Roboflow version training cancel** (Create). Cancel an in-progress Roboflow training job for a specific dataset version. The job must be in a running or queued state; cancelling a finished job returns 409 Conflict. Not idempotent. Returns: success. Required: workspace, project, version.
- **Create a Roboflow version training stop** (Create). Request an early stop for a running Roboflow training job on a specific project version. Unlike cancel, this operation is idempotent and returns success even if the job has already stopped. Returns: success. Required: workspace, project, version.
- **Create a Roboflow image** (Create). Upload an image to a Roboflow project dataset via URL or raw image body. Optionally assign the image to a split, batch, or tag sequence. Returns: success, taskId, signedUrl, url. Required: project.
- **Get single Roboflow image by ID** (Get). Get details about a specific image in a Roboflow project by id. Returns the image object containing id, name, annotation (with bounding boxes), labels, split, tags, created, urls, and embedding. Required: workspace, project, id.
- **Delete a Roboflow image by ID** (Delete). Delete one or more images from a Roboflow project dataset by specifying their IDs in the request body. Returns an empty 204 response on success. Required: workspace, project, images.
- **Update a Roboflow image tag by ID** (Update). Add, remove, or set tags on a specific Roboflow image. Returns: success. Required: workspace, project, image_id, operation, tags.
- **Create a Roboflow annotation** (Create). Upload and attach an annotation file to an existing image in a roboflow dataset. Supports any roboflow-compatible annotation format; send raw annotation content as the request body, or a JSON body with annotationFile and labelmap for index-based formats like Darknet TXT. Returns an empty 204 response on success. Required: project, image_id, name.
- **Create a Roboflow dataset zip** (Create). Initiate an async dataset zip upload for a Roboflow project. Sends a POST that returns a signed GCS URL to PUT the zip file to and a task identifier for polling status. Returns: taskId, signedUrl, url. Required: project.
- **List all Roboflow batch** (List). List all image batches for a Roboflow project. Returns a collection of batch records each containing id, name, images, numJobs, and uploaded timestamp. Required: workspace, project.
- **Get single Roboflow batch by ID** (Get). Get a single Roboflow image batch by id within a project. Returns the batch's id, name, images, numJobs, and uploaded timestamp. Required: workspace, project, id.
- **Create a Roboflow annotation job** (Create). Create a new annotation job in a Roboflow project, assigning images from an upload batch to a labeler and reviewer. Returns: id, name, status, created, labeler, reviewer, numImages, annotated, unannotated, approved, rejected, owner, createdBy, instructionsText, sourceBatch, project. Required: workspace, project, name, batch, num_images, labelerEmail, reviewerEmail.
- **List all Roboflow annotation job** (List). List all annotation jobs in a Roboflow project. Returns each job's id, name, status, labeler, reviewer, numImages, annotated, unannotated, approved, rejected, sourceBatch, created, instructionsText. Required: workspace, project.
- **Get single Roboflow annotation job by ID** (Get). Get a single Roboflow annotation job by id. Returns the full job object including id, name, status, labeler, reviewer, numImages, annotated, unannotated, approved, rejected, sourceBatch, instructionsText, createdBy, owner, project, created. Required: workspace, project, id.
- **Create a Roboflow auto label job** (Create). Create a Roboflow auto-label job to automatically annotate a batch of images in a project using SAM 3 or a custom Roboflow model. Returns: jobId, annotationJobId, message. Required: workspace, project, batchId, modelType.
- **Get single Roboflow auto label job by ID** (Get). Get the status and progress details of a Roboflow auto-label job by id. Returns: id, status, annotationJobId, jobId, message. Required: workspace, id.
- **List all Roboflow annotation insight** (List). List annotation insights for a Roboflow workspace. Returns per-labeler metrics in data (approved, boxesDrawn, imagesLabeled, approvalRate, labelerId, projectId), labeler metadata in labelers (id, displayName, email), and aggregated workspace-level totals in stats (numImagesLabeled, numBoxesDrawn, approvalRate). Required: workspace, startDate, endDate. Data is only available from 2023-08-01 onward.
- **List all Roboflow workflow** (List). List all Roboflow workflows in a workspace. Returns: id, name, url, createdAt, workflows, status. Required: workspace.
- **Get single Roboflow workflow by ID** (Get). Get a single Roboflow workflow by id. Returns the workflow metadata including id, name, url, and createdAt. Required: workspace, id.
- **Delete a Roboflow workflow by ID** (Delete). Soft-delete (move to Trash) a Roboflow workflow by id. Returns: deleted, workspace, workflow, workflowId, trash, workflows, status. Required: workspace, id.
- **Create a Roboflow workflow** (Create). Create a new Roboflow workflow in a workspace. Returns the created workflow's id and url. Required: workspace.
- **Update a Roboflow workflow by ID** (Update). Update a Roboflow workflow's name, URL slug, or specification. Returns the updated workflow's id and url. Required: workspace, id.
- **Create a Roboflow workflow fork** (Create). Fork a Roboflow workflow from another workspace into the specified target workspace. Returns the forked workflow object including its id and url, plus a status indicator. Required: workspace, source_workspace, source_workflow.
- **List all Roboflow workflow version** (List). List versioned snapshots of a Roboflow workflow specification. Returns version records including id. Required: workspace, workflow_url.
- **Create a Roboflow workflow token** (Create). Generate a short-lived Roboflow workflow token suitable for executing workflows from a public client such as a browser. Returns: status, workflows. Required: workspace.
- **Create a Roboflow project folder** (Create). Create a new project folder in a Roboflow workspace to organize projects. Returns: id. Required: workspace.
- **Get single Roboflow project folder by ID** (Get). Get a Roboflow project folder by id. Returns the ProjectFolder object including its id, name, projects, external_id, and auth_groups. Required: workspace, id.
- **Update a Roboflow project folder by ID** (Update). Update a Roboflow project folder's properties by id. Returns an empty 204 response on success; pass returnUpdated=true as a query parameter to receive the updated folder data in the response instead. Required: workspace, id.
- **Delete a Roboflow project folder by ID** (Delete). Delete a Roboflow project folder by id. All projects within the folder are placed back into the top-level workspace and not deleted. Returns an empty 204 response on success. Required: workspace, id.
- **Update a Roboflow project folder project by ID** (Update). Add one or more projects to an existing Roboflow project folder. Returns an empty 204 response on success. Required: workspace, folder_id, projects.
- **Delete a Roboflow project folder project by ID** (Delete). Remove one or more projects from a Roboflow project folder, placing them back into the top-level workspace. Returns an empty 204 response on success. Required: workspace, folder_id, projects.
- **List all Roboflow trash** (List). List all items currently in a Roboflow workspace's Trash, including projects, versions, and workflows. Returns: sections, deleted, workspace, project, projectId, trash. Required: workspace.
- **Create a Roboflow trash restore** (Create). Restore a Roboflow project, version, or workflow from Trash. Returns: deleted, workspace, project, projectId, version, workflow, workflowId, trash. Required: workspace, type, id. For version restores, parentId (parent project ID) is also required and the parent project must be active (not itself in Trash).
- **Get single Roboflow async task by ID** (Get). Get the status of a roboflow async task by id. Returns: taskId, status, progress, result, url. Required: workspace, id.
- **List all Roboflow API key** (List). List all Roboflow API keys for a workspace, returning masked key metadata and the workspace publishable key. Returns: apiKeys, publishableKey, apiKey. Required: workspace.
- **Get single Roboflow API key by ID** (Get). Get masked metadata for a single Roboflow API key by its id. Returns the apiKey object containing keyId, name, prefix, scopes, folderIds, default, protected, and disabled. Required: workspace, id.
- **Create a Roboflow API key** (Create). Create a new Roboflow API key in a workspace. The secret key value is returned only once in the key field — store it securely immediately. Returns: keyId, key, name, scopes, folderIds, protected, publishableKey, apiKey. Required: workspace.
- **Update a Roboflow API key by ID** (Update). Update a Roboflow API key's name, scopes, metadata, or protected/disabled state; send only the fields you want to change. Returns the updated apiKey object with keyId, name, prefix, scopes, folderIds, protected, and disabled. Required: workspace, id.
- **Delete a Roboflow API key by ID** (Delete). Revoke a Roboflow API key permanently by its id; existing applications using it will immediately fail to authenticate. Returns: status, keyId, apiKey. Required: workspace, id.
- **List all Roboflow publishable key** (List). Get the publishable key for a Roboflow workspace. The key is a permanent, non-secret credential of the form `rf_<workspaceId>` that can be safely embedded in client-side code for inference and model downloads. Returns: publishableKey, status, keyId. Required: workspace.
- **Create a Roboflow agent chat** (Create). Send a natural-language message to the Roboflow AI agent to create or edit Workflows in a workspace. Returns: text, workflows, conversation_id, workflowId, workflowUrl, versionId, status. Required: workspace, message. Pass conversation_id to continue an existing conversation; omit it to start a new one. Workflow changes are saved as drafts.
- **Create a Roboflow agent workflow publish** (Create). Publish the latest draft version of a Roboflow agent Workflow, deploying it to production. Returns: workflowId, workflowUrl, versionId, status. Required: workspace, workflow_url. Returns 400 if there is no unpublished draft to deploy.
- **List all Roboflow agent conversation** (List). List all Roboflow agent conversations in a workspace. Returns: id, workflowId, workflowUrl, versionId, status. Required: workspace. Optionally filter by source (api or web) or by a workflow URL slug.
- **Get single Roboflow agent conversation by ID** (Get). Get a single Roboflow agent conversation by id, including its full message history. Returns: id, workflowId, workflowUrl, versionId, status. Required: workspace, id.
- **List all Roboflow project deploy** (List). Get the Roboflow project deployment state for a given workspace and project. Returns: project, workflow, model, deployability, activeLearning, baseWorkflowWasCreated. Required: workspace, project.
- **Update a Roboflow project deploy model by ID** (Update). Update the deployed model for a roboflow project, setting the model used by the project's deployment workflow. Accepts three model shapes: a specific model by ID (`model_id`), zero-shot segmentation (`sam3`), or zero-shot classification (`clip`). Returns the updated deployment state including project, workflow, model, deployability, activeLearning, and baseWorkflowWasCreated. Required: workspace,…
- **List all Roboflow active learning** (List). List the Active Learning configuration for a roboflow project, returning its enabled state, collection limits, filters, deployment readiness, model settings, and workflow details. Returns: project, workflow, model, deployability, activeLearning, baseWorkflowWasCreated. Required: workspace, project.
- **Create a Roboflow active learning enable** (Create). Enable Active Learning on a Roboflow project. If no model is configured, the service attempts to auto-select one. Returns: project, workflow, model, deployability, activeLearning, baseWorkflowWasCreated. Returns 400 if the workflow is not deployable. Required: workspace, project.
- **Create a Roboflow active learning disable** (Create). Disable Active Learning on a Roboflow project. Returns the updated deployment configuration including project, workflow, model, deployability, and activeLearning (with enabled set to false). Required: workspace, project.
- **Update a Roboflow active learning config by ID** (Update). Update the Active Learning configuration for a Roboflow project's deployment, setting collection limits (data percentage, rate limits, batch frequency, image sizing, tags) and optional image filters. Returns the updated deployment state including project, workflow, model, deployability, and activeLearning objects. Required: workspace,id.
- **List all Roboflow active learning image** (List). List Active Learning images in the Roboflow project's review queue with optional filtering and sorting. Returns: activeLearningImages, project, workflow, model, deployability, activeLearning, baseWorkflowWasCreated. Required: workspace, project.
- **List all Roboflow credential** (List). List credentials in a roboflow workspace with their secrets masked. Returns a collection of credential records including id, owner, type, name, and payload (with secrets masked). Required: workspace.
- **Create a Roboflow credential** (Create). Create a credential in a Roboflow workspace to store cloud access secrets (AWS IAM, GCS, API key, etc.). Returns: id. Required: workspace, type, name, payload.
- **Delete a Roboflow credential by ID** (Delete). Delete a credential from a roboflow workspace by id. Returns an empty 204 response on success. Required: workspace, id.
- **List all Roboflow bucket mirror config** (List). List roboflow bucket mirror configurations in a workspace, each with a summary of its latest sync job. Returns: id, name, latestJob. Required: workspace.
- **Create a Roboflow bucket mirror config** (Create). Create a roboflow bucket mirror configuration to mirror files from an external S3 or GCS bucket into a workspace. Returns: id. Required: workspace, name, credentialId, bucket, mirrorConfigs.
- **Get single Roboflow bucket mirror config by ID** (Get). Get a single roboflow bucket mirror configuration by id. Returns: id, name, latestJob. Required: workspace, id.
- **Update a Roboflow bucket mirror config by ID** (Update). Update a roboflow bucket mirror configuration by id. Send only the fields you want to change. Returns: id, name. Required: workspace, id.
- **Delete a Roboflow bucket mirror config by ID** (Delete). Delete a roboflow bucket mirror configuration by id. Returns an empty 204 response on success. Required: workspace, id.
- **List all Roboflow bucket mirror validation** (List). Validate a Roboflow bucket mirror configuration by checking that the referenced credential can list, head, and read objects in the configured bucket. Returns: data. Required: workspace, config_id.
- **Create a Roboflow bucket mirror sync** (Create). Trigger a Roboflow bucket mirror sync for a given configuration. Validates the configuration first, then starts a mirror run and returns one batch id per mirror rule. Manual runs are rate limited. Returns: id. Required: workspace, config_id.
- **Get single Roboflow bucket mirror job by ID** (Get). Get a single Roboflow bucket mirror sync job by its id, allowing you to poll job progress after triggering a run. Returns: batchId, status, id. Required: workspace, config_id, id.
- **Create a Roboflow vision event** (Create). Create a single vision event in roboflow to record an observation from a computer vision deployment. Returns: eventId, created. Required: eventId, eventType, useCaseId, timestamp, eventData. Duplicate eventId values overwrite previously ingested events.
- **Create a Roboflow vision event batch** (Create). Batch-create up to 100 Roboflow vision events in a single request, more efficiently than creating events individually. Returns: created, eventIds, deprecations. Required: events. Maximum of 100 events per batch.
- **Create a Roboflow vision event image** (Create). Upload an image to Roboflow for use in vision events. This is a workspace-level upload not tied to any specific project; the returned sourceId can be referenced in the images array when creating vision events. Returns: success, sourceId, url. Required: file. Supported formats: JPEG, PNG, WebP, GIF, TIFF, BMP, HEIC, and AVIF.
- **Create a Roboflow vision event query** (Create). Query roboflow vision events with filters for time range, event type, device context, detection classes, custom metadata, and event-specific fields. Returns: error. Required: useCaseId. Max 1000 events per page; defaults to 100.
- **List all Roboflow vision event metadata schema** (List). Get the custom metadata schema for Roboflow vision events for a given use case, returning all observed field names and their data types across ingested events. Returns: created, eventIds. Required: useCaseId.
- **List all Roboflow use case** (List). List all Roboflow use cases in your workspace that have recorded vision events. Returns: id, name, status, workspaceId, createdAt, updatedAt, error. Optionally filter by use case status (active or inactive).
- **Create a Roboflow use case** (Create). Create a new use case in your Roboflow workspace to organize vision events by deployment context. Returns: id, name. Required: name.
- **Update a Roboflow use case by ID** (Update). Update the name or status of an existing Roboflow use case. Returns: id, name. Required: id. At least one of name or status must be provided in the request body.
- **Create a Roboflow use case archive** (Create). Archive a roboflow use case by setting its status to inactive, hiding it from active listings and preventing new event ingestion (soft delete — data is preserved). Returns: success, error. Required: use_case_id.
- **Create a Roboflow use case unarchive** (Create). Restore a previously archived Roboflow use case by setting its status back to active. Only use cases with an inactive status can be unarchived. Returns: success, error. Required: use_case_id.
- **List all Roboflow dedicated deployment machine type** (List). List all available machine types for Roboflow dedicated deployments. Returns: name, description, deployment_name, status, machine_type, public_url, created_at, expires_at.
- **Create a Roboflow dedicated deployment** (Create). Create a Roboflow dedicated deployment (managed GPU machine). Provisions asynchronously — poll GET /get until status is 'ready'. Returns: deployment_name, machine_type, status, public_url, created_at, expires_at. Required: deployment_name, machine_type, creator_email, duration, delete_on_expiration.
- **List all Roboflow dedicated deployment** (List). List all Roboflow dedicated deployments in the workspace. Returns records with deployment_id, deployment_name, machine_type, status, creator_email, subdomain, domain, duration, inference_version, min_replicas, max_replicas, num_replicas, workspace_id, workspace_url.
- **Delete a Roboflow dedicated deployment by ID** (Delete). Delete a Roboflow-dedicated deployment by ID. Returns: message, deployment_name, status, machine_type, public_url, created_at, expires_at. Required:id.
- **Get single Roboflow dedicated deployment by ID** (Get). Get a single Roboflow dedicated deployment by id. Returns: deployment_name, machine_type, status, workspace_url, public_url, created_at, expires_at. Required:id.
- **Create a Roboflow dedicated deployment pause** (Create). Pause a Roboflow dedicated deployment, stopping the running GPU machine while preserving its configuration. Returns: message, deployment_name, status, machine_type, public_url, created_at, expires_at. Required: deployment_name.
- **Create a Roboflow dedicated deployment resume** (Create). Resume a paused Roboflow dedicated deployment. Returns: message, deployment_name, status, machine_type, public_url, created_at, expires_at. Required: deployment_name.
- **List all Roboflow dedicated deployment log** (List). List log entries for a Roboflow dedicated deployment, with optional timestamp range filtering and a cap on the number of returned entries. Returns: deployment_name, status, machine_type, public_url, created_at, expires_at. Required: deployment_name. Omit from_timestamp and to_timestamp to fetch the most recent entries up to max_entries.
- **List all Roboflow dedicated deployment usage workspace** (List). List workspace-wide usage data in Roboflow for a specified time window across all dedicated deployments. Returns: deployment_name, status, machine_type, public_url, created_at, expires_at. Required: from_timestamp, to_timestamp.
- **List all Roboflow dedicated deployment usage** (List). List per-deployment usage in Roboflow for a specific dedicated deployment within a given time window. Returns: deployment_name, status, machine_type, public_url, created_at, expires_at. Required: deployment_name, from_timestamp, to_timestamp.
- **Create a Roboflow inference model** (Create). Run inference on a Roboflow trained model by submitting an image to a specific model version. Returns predictions whose structure depends on the model type (object detection, classification, segmentation); consult the Roboflow deployment documentation for per-task response shapes. Required: workspace, project, version.
- **Create a Roboflow inference workflow** (Create). Run a Roboflow inference workflow by posting an inputs payload to the specified workspace and workflow. Returns workflow-specific inference results whose exact shape is determined by the workflow's configured output steps; consult the workflow definition in roboflow for the field-level breakdown. Required: workspace, workflow.