Artificial Intelligence
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.
-
How it connects. Connects with an API key. The credential goes into a vault nobody reads back.
-
One address. https://api.elaichi.ai/mcp, the same for every user.
-
Their own access. An agent never gets more than the person it acts for.
How to connect
How to connect Roboflow to Claude, ChatGPT or Cursor
Two steps, about a minute.
In Elaichi
Connect Roboflow once
-
Open Connections, choose Add connection, and pick Roboflow.
-
Optionally set Share with to give a team access, then press Connect.
-
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.
The credential is vaulted. Nobody reads it back, not even the AI.
Add connection
Choose a connector.
In your AI client
Point it at one endpoint
Everyone in the organization uses the same address, and each person only ever reaches what their own account allows.
Connect Roboflow to Claude
-
1
Open Customize, then Connectors.
-
2
Press Add.
-
3
Name it, paste the MCP server URL, then Continue.
https://api.elaichi.ai/mcp -
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
-
1
Open Plugins, then press the + button.
-
2
Name it and paste the endpoint into Server URL.
https://api.elaichi.ai/mcp -
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
-
1
Open
~/.cursor/mcp.json. -
2
Add the endpoint under
mcpServers.https://api.elaichi.ai/mcp -
3
Reload Cursor, then sign in and approve.
~/.cursor/mcp.json
{
"mcpServers": {
"elaichi": {
"url": "https://api.elaichi.ai/mcp"
}
}
}
Set up per machine, so repeat it on each computer you work from.
Connect Roboflow to any MCP client
-
1
Add the endpoint as a remote MCP server.
https://api.elaichi.ai/mcp -
2
Sign in and approve.
{
"mcpServers": {
"elaichi": {
"url": "https://api.elaichi.ai/mcp"
}
}
}
The Elaichi Agent already has these tools, with nothing to set up.
Use cases
What teams do with Roboflow through Elaichi
Every one of these runs inside the access the person already has, and lands in the same audit log.
-
Machine learning
Kick off a training run without opening a tab
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.
-
Data annotation
Upload new images into the right project
Send a batch of images straight into the Roboflow project they belong to, then confirm they landed by asking for them back by ID.
-
Computer vision engineering
Compare dataset versions before you ship
Pull up every version of a project, see which ones have exports and trained models, and pick the one to promote.
-
Operations
Check project health across the workspace
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.
-
Product
Fork a project to try a new idea
Copy an existing project into a fresh fork so a new detection experiment starts from real labeled data instead of an empty dataset.
-
Research
Keep the workspace tidy
List every project and version in a workspace, spot the abandoned experiments, and remove the ones nobody trained in months.
Try asking
- “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.”
AI tools
Roboflow tools for your AI agents
103 tools are ready to call through Elaichi's MCP endpoint the moment you connect Roboflow, governed by the same roles, restrictions, and audit log as everything else in Elaichi.
No tools match your search.
See it in Elaichi
What connecting Roboflow gets you
6 screens from the product, each doing one job for your Roboflow account.
The agent
Ask about projects, get real answers.
Type a question in plain language and read it straight from live Roboflow records.
- workspaces
- projects
- versions
- exports
Ask Elaichi to work across your apps.
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.
Also runs in Claude, ChatGPT or Cursor
MCP clients
Claude, ChatGPT and Cursor share one endpoint.
Connect Roboflow through one org-wide MCP endpoint over OAuth, no SDK, no shared key.
Copy the endpoint
Tool catalog
103 Roboflow tools, no custom code.
Workspaces, projects, versions and exports are already wired as callable MCP tools.
- List all Roboflow workspace
- Get single Roboflow workspace by ID
- Create a Roboflow workspace search
- Create a Roboflow project
Toolboxes
Every team gets its own toolbox.
Curate one Roboflow toolbox per team and scope it to the people who need it.
- Machine learning
- Data annotation
- MLOps
- Research
Shared connections
Teammates use Roboflow without seeing keys.
One person connects the account, teams and members work through it with no credential.
- Computer vision
- Annotation team
- Research lab
- Production models
Audit log
Every Roboflow call is on the record.
When, who, what happened, type and resource, appended for every project and version touched.
- When
- Who
- What happened
- Type
Launching soon
From answering questions to doing the work
A person no longer has to ask. A trigger starts the work, inside the same permissions and the same audit log as everything else. Automations and live dashboards are launching soon, on the Black plan.
Automations
A schedule fires, a digest ships.
Fetch Roboflow projects, group them, draft a digest, get approval, post it back.
Roboflow digest
Run 418 · started 2 minutes ago · on behalf of Emma Laurent
-
✓
Schedule
Every weekday at 08:00
0.2s -
✓
Fetch workspaces
Roboflow
1.4s -
✓
Group by owner
Transform
0.1s -
✓
Draft the digest
Agent step
Ran with 4 tools, returned a structured summary
6.2s -
Approve the digest
Needs approval
Assigned to Michael Brennan
Approve Deny -
Post the digest
Roboflow
Queued
Collections and dashboards
Roboflow project health, counted on schedule.
Four metrics, 14 days of records created and a team breakdown, computed without a model.
Roboflow health
Refreshed 4 minutes ago · every 15 minutes · from the workspaces collection
Workspaces
1,284 ↓ 12%
Projects
96 ↓ 8%
Needs attention
3 ↑ 2
Updated this week
412 ↑ 9%
Workspaces created
Last 14 days
By team
Share of activity
Computer vision 34%
Annotation team 27%
Research lab 21%
Production models 18%
FAQ
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 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.
Put Roboflow in front of your team
Fourteen days on Gold, no credit card. Connect it once and pick what each team can call.