Artificial Intelligence
OpenPipe MCP connector
The OpenPipe connector brings your logged calls, datasets, fine-tuned models and evaluation judges to Claude, ChatGPT, Cursor and any MCP client, so your team can work with OpenPipe from the tools they already use.
-
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.
23 tools
How to connect
How to connect OpenPipe to Claude, ChatGPT or Cursor
Two steps, about a minute.
In Elaichi
Connect OpenPipe once
-
Open Connections, choose Add connection, and pick OpenPipe.
-
Optionally set Share with to give a team access, then press Connect.
-
Paste an OpenPipe API key. One person generates a token in OpenPipe 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 OpenPipe 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 OpenPipe 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 OpenPipe 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 OpenPipe 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 OpenPipe through Elaichi
Every one of these runs inside the access the person already has, and lands in the same audit log.
-
AI engineering
Turn logged calls into a training dataset
Pull the latest logged calls from OpenPipe, tag the good ones, and add them as entries to a dataset without leaving the chat you are already working in.
-
AI engineering
Kick off a fine-tune from a conversation
Pick a dataset, choose a base model, and start a fine-tuning job in OpenPipe, then check on it later by asking how the run is going.
-
Data science
Keep datasets tidy before a training run
List every dataset in OpenPipe, spot the stale or duplicate ones, and remove them so the next fine-tune starts from a clean set.
-
Quality
Score model output against your own criteria
Run a criteria judge in OpenPipe on a batch of responses and get a plain summary of where the model passed and where it fell short.
-
Product
Label production calls by feature and customer
Update tags and metadata on logged calls in OpenPipe after a launch, so the team can filter later by the feature or account that produced them.
-
Platform
Compare the models you have deployed
List the models in OpenPipe, look up any single one by ID, and retire the ones nobody is calling anymore.
Try asking
- “List our OpenPipe datasets and their entry counts.”
- “Which OpenPipe models were called most this week.”
- “Show the latest logged calls and their tags.”
AI tools
OpenPipe tools for your AI agents
23 tools are ready to call through Elaichi's MCP endpoint the moment you connect OpenPipe, 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 OpenPipe gets you
6 screens from the product, each doing one job for your OpenPipe account.
The agent
Ask about datasets, get answers back.
Plain language questions run against live OpenPipe datasets, models and logged calls.
- datasets
- dataset entries
- models
- logged calls
Ask Elaichi to work across your apps.
List our OpenPipe datasets and their entry counts.
Which OpenPipe models were called most this week.
Show the latest logged calls and their tags.
Also runs in Claude, ChatGPT or Cursor
MCP clients
Claude, ChatGPT and Cursor share one endpoint.
Reach OpenPipe from any MCP client over OAuth, no SDK, no shared key.
ElaichiMCP clientsClaude
ChatGPT
Cursor
Copy the endpoint
Tool catalog
23 OpenPipe tools ready to call.
Datasets, dataset entries, models, chat completions and logs, with no custom code.
- Create a OpenPipe check cach
- Create a OpenPipe chat completion
- Create a OpenPipe report
- Create a OpenPipe report anthropic
ElaichiToolsToolboxes
Every team gets its own toolbox.
Curate one OpenPipe toolbox per team so each sees only relevant tools.
- Machine learning
- Data platform
- Product engineering
- Quality
ElaichiToolboxesShared connections
Teammates use OpenPipe without seeing keys.
One connected OpenPipe account, shared to named teams and members, credential never exposed.
- Machine learning
- Data platform
- Product engineering
- Quality
ElaichiConnectionsAudit log
Answer who touched which dataset.
Every OpenPipe call is logged with time, person, action, type and resource.
- When
- Who
- What happened
- Type
ElaichiAudit logLaunching 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.
Six steps fetch OpenPipe logged calls, group them, draft, get approval, then post back.
OpenPipe digest
Run 418 · started 2 minutes ago · on behalf of Emma Laurent
-
✓
Schedule
Every weekday at 08:00
0.2s -
✓
Fetch datasets
OpenPipe
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
OpenPipe
Queued
Collections and dashboards
OpenPipe numbers land before anyone asks.
Four metrics, 14 days of dataset entries and a team breakdown, computed on schedule.
OpenPipe health
Refreshed 4 minutes ago · every 15 minutes · from the datasets collection
Datasets
1,284 ↓ 12%
Dataset entries
96 ↓ 8%
Needs attention
3 ↑ 2
Updated this week
412 ↑ 9%
Datasets created
Last 14 days
By team
Share of activity
Machine learning 34%
Data platform 27%
Product engineering 21%
Quality 18%
FAQ
Frequently asked questions
How do I connect OpenPipe to Claude?
Connect OpenPipe in Elaichi first, which means pasting the API key you copy from your OpenPipe project settings. There is no OAuth application to register and no client ID or secret to generate. Then open Claude, go to Customize, then Connectors, then Add, and paste the endpoint https://api.elaichi.ai/mcp. Sign in as yourself and OpenPipe is available in your next conversation.
Does OpenPipe work with ChatGPT and Cursor as well as Claude?
Yes. Once OpenPipe is connected in Elaichi, the same endpoint, https://api.elaichi.ai/mcp, works in Claude, ChatGPT, Cursor, any other MCP client and the Elaichi Agent. You connect OpenPipe once and every client picks it up.
What can an AI agent actually do with my OpenPipe data?
An agent can fetch your latest logged calls from OpenPipe, tag them, update their metadata and add the useful ones to a dataset. It can create and remove datasets, start a fine-tuning job, check how a fine-tune is going, list and look up your models, run a criteria judge on responses and send chat completions through OpenPipe. In short, the everyday work of building a dataset and training a model, done by asking.
Does connecting OpenPipe give the AI access to every project?
No. Access follows the OpenPipe API key the person connected, so the agent can only reach the project and records that key already covers. Elaichi can narrow that further by turning off individual actions, but it can never widen access beyond what OpenPipe itself grants.
Can I stop an agent from deleting datasets or models in OpenPipe?
Yes. Restrictions in Elaichi work per action, so you can allow an agent to read logged calls and add dataset entries while blocking it from deleting datasets, models or fine-tuning jobs in OpenPipe. A blocked action is never advertised to Claude, ChatGPT or any other client, so no prompt can reach it.
What happens to an OpenPipe connection when someone leaves?
Offboarding that person in Elaichi ends their access to OpenPipe through every client at once. If they shared an OpenPipe connection with a team, it keeps working for everyone else on that team. Disconnecting OpenPipe once in Elaichi removes it from Claude, ChatGPT, Cursor and every other client at the same time.
Put OpenPipe in front of your team
Fourteen days on Gold, no credit card. Connect it once and pick what each team can call.