Humanloop MCP for AI. Control Prompt Versions & Deployment Pipeline

Q: Can I see the full version history of a specific prompt?

Yes! Use the listpromptversions tool with the Prompt ID. It will return all historical versions, allowing you to track changes and metadata over time.

Q: How do I deploy a prompt version to a specific environment like production?

You can use the deployprompt tool. Provide the Prompt ID and the Environment ID to set that specific version as the active deployment for that environment.

Q: Is it possible to record model outputs for later evaluation?

Absolutely. Use the logtoprompt tool to record a generation, including the prompt path, messages, and output, which can then be used for evaluation in Humanloop.

Q: How do I list all prompts using listprompts if my organization has many entries?

The listprompts tool returns all available prompt IDs and basic names. If you have thousands of prompts, you might need to implement pagination in your agent logic to ensure you retrieve every single entry.

Q: What is the best way to update a prompt's name using updatepromptversion?

You must use updatepromptversion and provide both the version ID and the new desired metadata. This prevents accidental changes, so always confirm you have the correct unique version identifier first.

Q: If I need to remove an old prompt using deletepromptversion, are there any dependencies?

Deleting a prompt version is irreversible. Before running deletepromptversion, verify that no active deployments or critical logs still reference that specific version ID.

Q: Can I check the status of all environments and their deployed versions using listpromptenvironments?

Yes, listpromptenvironments gives a clear overview. It shows every configured environment (like staging or production) alongside which prompt version is currently active in that location.

Q: Is there a way to create a new prompt configuration using upsertprompt if I don't have an ID?

The upsertprompt tool handles creation or updating. If no unique identifier is provided, the API will generate one for you and save the entire configuration set.

Claude

ChatGPT

Cursor

Gemini

Windsurf

VS Code

JetBrains

Vercel

See Vinkius in Action

Works with every AI agent you already use

…and any MCP-compatible client

How this MCP server connects to your AI agent

Humanloop (LLM Prompt Management API) MCP. Manage your entire prompt lifecycle—versioning, deployment, and monitoring—directly from any AI agent interface. You can track model generations, set active versions for staging or production, and audit prompt history without leaving your editor.

It's the control plane you need for reliable LLMOps.

What AI agents can do with Humanloop (LLM Prompt Management API) Automation

Call prompt stream

Executes a specified prompt and streams the generated text response back to the user's client.

Delete prompt version

Permanently removes an identified version of a stored prompt from the library.

Deploy prompt

Sets a specific, targeted prompt version as the active configuration for one or more designated environments.

+ 9 more capabilities included

List all prompts

Retrieves a complete inventory of every prompt stored in the organization.

View and update versions

Accesses version history for specific prompts, allowing you to modify metadata or remove outdated copies.

Deploy active prompt versions

Sets a targeted prompt version as the live configuration for defined environments like production or staging.

Execute and stream prompts

Runs a prompt and streams the resulting text back to your agent client in real-time.

Log model generations

Records detailed logs for any model generation run against a specific prompt, aiding evaluation.

Manage deployment settings

Adds or removes version deployments from specific environments and updates monitoring rules.

Ask an AI about this

Included with Plan

Waiting for input…

AI Agent

What AI agents can do with Humanloop (LLM Prompt Management API) with 12 Tools

These twelve tools give your agent everything it needs to manage the full lifecycle of prompts, from initial creation and versioning through deployment and monitoring.

Make your AI actually useful.

Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.

Start using Humanloop (LLM Prompt Management API) on Vinkius

Call Prompt Stream

Executes a specified prompt and streams the generated text response back to the user's client.

Delete Prompt Version

Permanently removes an identified version of a stored prompt from the library.

Deploy Prompt

Sets a specific, targeted prompt version as the active configuration for one or more...

Get Prompt

Retrieves all detailed information about a single prompt using its unique ID.

List Prompt Environments

Lists all existing environments and shows which version of each prompt is currently...

List Prompt Versions

Gathers a complete list of every saved historical version associated with one specific prompt.

List Prompts

Retrieves a full directory listing of all prompts currently stored in the organization's library.

Log To Prompt

Creates a record (a log) detailing model outputs for a specific prompt ID, useful...

Remove Deployment

Deactivates and removes a previously deployed version from a specified environment.

Update Monitoring

Activates or deactivates monitoring evaluators for logs associated with a prompt.

Update Prompt Version

Changes the descriptive name or metadata of an existing, specific prompt version.

Upsert Prompt

Creates a brand new prompt configuration or updates an existing one with new content.

Security and governance baked right in.

Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.

Claude AI

Open Claude Settings

Go to claude.ai, click your profile icon, then navigate to Customize → Connectors.

Add Custom Connector

Click the "+" button and select Add custom connector. Paste your Vinkius endpoint URL:

https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp

Replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com. For OAuth-protected servers, expand Advanced settings to add credentials.

Start a conversation

Open a new chat. The Humanloop integration is available immediately — no restart needed.

Antigravity

Configure Agent Environment

Open your Antigravity agent's workspace configuration or mcp-servers.json file.

Bind the Endpoint

Add the Vinkius endpoint URL to your agent's MCP connections list:

"mcp_servers": {
  "humanloop-llm-prompt-management-api": {
    "serverUrl": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
  }
}

Provide your secure token in place of [YOUR_TOKEN_HERE] to ensure your agent requests are authenticated.

Execute

Start your Antigravity session. The agent will autonomously discover and utilize the Humanloop tools with full Vinkius guardrails applied.

VS Code Copilot

⚡

One-Click Install (Recommended)

In your Vinkius Dashboard, simply click the Add to VS Code button for this server. We'll automatically configure your local workspace.

Or configure manually

Open MCP Settings

Open VS Code, press Ctrl/Cmd + Shift + P, and search for GitHub Copilot: MCP Servers.

Add Server Config

Add the Vinkius endpoint configuration to your mcp-servers.json file:

"humanloop-llm-prompt-management-api": {
  "url": "https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp"
}

Ensure you replace [YOUR_TOKEN_HERE] with your token from cloud.vinkius.com.

LangChain

Install Dependencies

Install the LangChain MCP adapters for your environment:

pip install langchain-mcp-adapters

Connect the Server

Use the SSEClient in LangChain to connect to the Vinkius managed endpoint:

from langchain_mcp_adapters.client import SSEClient

# Connect to Vinkius
client = SSEClient(url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")
tools = client.get_tools()

CrewAI

Define the Tool

Load the Vinkius MCP tools into your CrewAI agents:

from crewai import Agent
from mcp_crewai import MCPTool

# Connect securely to Vinkius
vinkius_tools = MCPTool(url="https://edge.vinkius.com/[YOUR_TOKEN_HERE]/mcp")

# Assign to Agent
researcher = Agent(
    role='Data Researcher',
    tools=vinkius_tools.get_all()
)

Execute Task

Run your CrewAI process. The agent will autonomously route tasks to the Vinkius managed server.

Choose How to Get Started

Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.

Build Your Own

Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.

Import from OpenAPI, Swagger, or YAML specs
Create Agent Skills with progressive disclosure
Deploy to edge with MCPFusion framework
Built in DLP, auth, and compliance on every call
Real time usage dashboard and cost metering
Publish to catalog or keep private

Start building

Make Your AI Do More

Start with Humanloop (LLM Prompt Management API), then connect any of our 5,100+ other servers whenever your AI needs more. One click, no limits.

Use this MCP plus 5,100+ others, all in one place
Add new capabilities to your AI anytime you want
Every connection is secured and compliant automatically
Track usage and costs across all your servers
Works with Claude, ChatGPT, Cursor, and more
New servers added to the catalog every week

Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Humanloop. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.

VINKIUS INFRASTRUCTURE

Cloud Hosted

Managed infra

V8 Isolated

Sandboxed per request

Zero-Trust Proxy

No stored credentials

DLP Enforced

Policy on every call

GDPR Compliant

EU data residency

Token Compression

~60% cost reduction

Your data is protected. See how we built it.

Built on the Model Context Protocol (MCP) for Claude, ChatGPT, Cursor, and more

The Model Context Protocol standardizes how applications expose capabilities to LLMs. Instead of operating in isolation, your AI gains direct access to external platforms, live data, and real-world actions through secure, standardized connections.

This connection provides 12 powerful capabilities that interface natively with Claude, ChatGPT, Cursor, and other compatible AI platforms. No middleware. No custom integration required.

The Headache of Prompt Drift, Solved with Vinkius AI Gateway

Right now, changing a single prompt requires a mess of clicks. You update the file in Git, then copy the new content into the staging dashboard, then manually test it in the UAT environment, and finally, you ask Ops to 'deploy it'—a process that takes half a day and generates at least three tickets.

With this MCP, you treat prompt management like any other code change. You manage versions and deployments directly via your agent client. It turns what was a multi-day coordination chore into a few commands.

The Humanloop (LLM Prompt Management API)

You skip the manual steps entirely. No more copy-pasting prompt JSONs between dashboards or emailing screenshots of versions. You just tell your agent what needs to change, and it handles the versioning and deployment status.

What's different now? Everything is traceable. Every prompt state, every environment setting, and every model output log lives in one governed system.

Support 24/7 support@vinkius.com ↗

Security Vinkius Trust Center ↗

SLA Service Level Agreement ↗

Report Listing Send Report ↗

prompt-engineering

llmops

prompt-management

ai-deployment

version-control

What your AI can actually do with this

This MCP lets you manage all your core prompts in one place. When you build an application using large language models, prompts are critical; they define how the AI behaves. Instead of relying on scattered files or remembering which prompt version is live, this connector gives you a central system to control them all.

You can list every prompt in your organization and check its full history. Need to test a new approach? Update a version's metadata or even delete obsolete copies to keep things clean. Once you find the right setup, you deploy it—setting specific versions as active for environments like staging or production with a simple command.

Plus, you can record model generations for later evaluation and monitor settings to ensure your LLM pipeline is healthy. Vinkius makes connecting this control layer easy; just plug into your preferred AI client.

Built · Hosted · Managed by Vinkius Humanloop MCP - Version & Deploy Prompts

Server ID 019e5d24-6db4-7286-8df7-beeb05480b28

Vinkius Inspector

Compliance Grade A+

Score 100/100

Report View Report ↗

Here's how it actually works

The bottom line is, your AI agent treats prompt management like it's built-in—you just talk to it instead of clicking through dashboards.

Subscribe to this MCP and enter your Humanloop API Key.

Your agent client connects, giving you read/write access to the prompt library.

You can then call functions like list_prompts or deploy_prompt directly from a conversation thread.

What Changes When You Connect

You never lose track of a prompt change. Use list_prompt_versions to see the full history, so you always know exactly which version is running in production.

Testing new prompts doesn't require manual deployment steps. You can use upsert_prompt and then test it via call_prompt_stream right away.

When a prompt breaks, don't panic. Use remove_deployment to instantly deactivate the bad version and roll back to a known good state.

Keep your models accountable by using log_to_prompt. This records every run, letting you audit performance later for quality assurance.

Manage complex environments easily. Run list_prompt_environments to see what's active across staging, production, and other targets at a glance.

See it in action

01 01

The QA team needs to audit the 'Support Bot' performance.

Instead of pulling logs from three different services, the agent runs list_prompts to find the ID. It then calls log_to_prompt repeatedly with specific inputs to capture model generations for a full quality review.

02 02

The DevOps team needs to roll out an urgent fix.

They use get_prompt to verify the current structure, then call update_prompt_version to stage the fix. Finally, they execute deploy_prompt to make it live in production instantly.

03 03

A PM needs to see what's running everywhere.

The agent calls list_prompt_environments. This immediately shows which specific prompt version is active across staging, the development sandbox, and the main client environment.

The honest tradeoffs

Manually tracking versions

Anti-pattern

A developer changes a prompt file locally, then manually tells the Ops team to 'check it out' without any version number or record.

The Fix

Always use list_prompts first. When ready for deployment, run update_prompt_version to formally tag the change and then call deploy_prompt to guarantee the version is immutable.

Overwriting current prompts

Anti-pattern

A developer calls a simple update function without checking if an environment is using that prompt, potentially breaking production instantly.

The Fix

Before updating anything, run list_prompt_environments to check dependencies. If the change breaks things, use remove_deployment on the old version before running any new deployment commands.

Ignoring historical data

Anti-pattern

Relying only on current chat logs to diagnose a prompt failure; you can't tell if the issue is recent or a pattern from weeks ago.

The Fix

Use log_to_prompt immediately after testing. This captures the necessary model generation data, allowing deep analysis even after the test session ends.

When It Fits, When It Doesn't

Use this MCP if your LLM prompt system is mission-critical—that means failure to deploy a correct prompt breaks revenue or core functionality. The focus here must be on governance: versioning and reliable deployment pipelines. Don't use it if you just need a simple, one-off chat bot; for that, basic API calls are fine. You should avoid using this MCP if your primary goal is just drafting content—that’s what the LLM client handles natively. However, you must use list_prompt_versions and deploy_prompt whenever you move a prompt from development to any environment other than local testing.

Questions you might have

Can I see the full version history of a specific prompt? +

Yes! Use the list_prompt_versions tool with the Prompt ID. It will return all historical versions, allowing you to track changes and metadata over time.

How do I deploy a prompt version to a specific environment like production? +

You can use the deploy_prompt tool. Provide the Prompt ID and the Environment ID to set that specific version as the active deployment for that environment.

Is it possible to record model outputs for later evaluation? +

Absolutely. Use the log_to_prompt tool to record a generation, including the prompt path, messages, and output, which can then be used for evaluation in Humanloop.

How do I list all prompts using `list_prompts` if my organization has many entries? +

The list_prompts tool returns all available prompt IDs and basic names. If you have thousands of prompts, you might need to implement pagination in your agent logic to ensure you retrieve every single entry.

What is the best way to update a prompt's name using `update_prompt_version`? +

You must use update_prompt_version and provide both the version ID and the new desired metadata. This prevents accidental changes, so always confirm you have the correct unique version identifier first.

If I need to remove an old prompt using `delete_prompt_version`, are there any dependencies? +

Deleting a prompt version is irreversible. Before running delete_prompt_version, verify that no active deployments or critical logs still reference that specific version ID.

Can I check the status of all environments and their deployed versions using `list_prompt_environments`? +

Yes, list_prompt_environments gives a clear overview. It shows every configured environment (like staging or production) alongside which prompt version is currently active in that location.

Is there a way to create a new prompt configuration using `upsert_prompt` if I don't have an ID? +

The upsert_prompt tool handles creation or updating. If no unique identifier is provided, the API will generate one for you and save the entire configuration set.

How this MCP server connects to your AI agent

What AI agents can do with Humanloop (LLM Prompt Management API) Automation

Call prompt stream

Delete prompt version

Deploy prompt

What AI agents can do with Humanloop (LLM Prompt Management API) with 12 Tools

Call Prompt Stream

Executes a specified prompt and streams the generated text response back to the user's client.

Delete Prompt Version

Permanently removes an identified version of a stored prompt from the library.

Deploy Prompt

Sets a specific, targeted prompt version as the active configuration for one or more...

Get Prompt

Retrieves all detailed information about a single prompt using its unique ID.

List Prompt Environments

Lists all existing environments and shows which version of each prompt is currently...

List Prompt Versions

Gathers a complete list of every saved historical version associated with one specific prompt.

List Prompts

Retrieves a full directory listing of all prompts currently stored in the organization's library.

Log To Prompt

Creates a record (a log) detailing model outputs for a specific prompt ID, useful...

Remove Deployment

Deactivates and removes a previously deployed version from a specified environment.

Update Monitoring

Activates or deactivates monitoring evaluators for logs associated with a prompt.

Update Prompt Version

Changes the descriptive name or metadata of an existing, specific prompt version.

Upsert Prompt

Creates a brand new prompt configuration or updates an existing one with new content.

Security and governance baked right in.

Claude AI

Open Claude Settings

Add Custom Connector

Start a conversation

Claude Code

Open your terminal

Add the MCP Server

Start coding

Cursor

One-Click Install (Recommended)

Open Cursor Settings

Add New Server

Use in Composer

Antigravity

Configure Agent Environment

Bind the Endpoint

Execute

VS Code Copilot

One-Click Install (Recommended)

Open MCP Settings

Add Server Config

Windsurf

One-Click Install (Recommended)

Open Windsurf Settings

Add Server Endpoint

LangChain

Install Dependencies

Connect the Server

CrewAI

Define the Tool

Execute Task

Choose How to Get Started

Build Your Own

Make Your AI Do More

Built on the Model Context Protocol (MCP) for Claude, ChatGPT, Cursor, and more

The Headache of Prompt Drift, Solved with Vinkius AI Gateway

The Humanloop (LLM Prompt Management API)

prompt-engineering

llmops

prompt-management

ai-deployment

version-control

What your AI can actually do with this

Here's how it actually works

Who is this actually for?

What Changes When You Connect

The QA team needs to audit the 'Support Bot' performance.

The DevOps team needs to roll out an urgent fix.

A PM needs to see what's running everywhere.