# meandr.com

[Meandr](https://meandr.com/)

[
#### Overview

One AI agent control plane to host, connect and govern your MCP servers.

/platform/ →](https://meandr.com/platform/)[
#### MCP Gateway

One workspace connection URL for every server, with routing and rate limits.

/mcp-gateway/ →](https://meandr.com/mcp-gateway/)[
#### MCP Hub

One team catalog of MCP servers and their tools.

/mcp-hub/ →](https://meandr.com/mcp-hub/)[
#### MCP Hosting

Run any MCP server on machines you choose, from a public catalog or your own container.

/mcp-hosting/ →](https://meandr.com/mcp-hosting/)[
#### MCP Governance

MCP policies allow, deny or hold each tool call for a person.

/mcp-governance/ →](https://meandr.com/mcp-governance/)[
#### MCP Security

Meandr signs in to your servers, so agents never hold the credentials.

/mcp-security/ →](https://meandr.com/mcp-security/)[
#### AI Agent Security

A key, a tool list and limits for each agent.

/ai-agent-security/ →](https://meandr.com/ai-agent-security/)[
#### AI Governance

Team roles, who approves held calls, and a record for audit.

/ai-governance/ →](https://meandr.com/ai-governance/)

[◆ Connect AWS MCP →](https://meandr.com/solutions/aws-mcp/)
MCP support:[Protocol versions](https://meandr.com/mcp-protocol-versions/)[Elicitation](https://meandr.com/mcp-elicitation/)
[Security & Trust →](https://meandr.com/security/)

**BY USE CASE**

[All solutionsThe overview — every use case and integration in one place.](https://meandr.com/solutions/)[Prevent dangerous callsAgents can't delete your data, and risky actions wait for your approval.](https://meandr.com/solutions/mcp-guardrails/)[Monitoring & auditEvery call on the record: agent, tool, decision, who approved.](https://meandr.com/solutions/audit/)[Hosted agent workflowsYour stdio MCP servers for ChatGPT and cloud agents, local or hosted.](https://meandr.com/solutions/local-mcp-server-remote-agents/)[Multi-agent systemsEach agent gets its own key, tools and call limit.](https://meandr.com/solutions/multi-agent-systems/)

**BY INTEGRATION**

[AWSAWS MCP (SigV4)signed for you, no proxy](https://meandr.com/solutions/aws-mcp/)[SStripe MCP refunds with human approvalrefunds over $500 wait for finance](https://meandr.com/solutions/stripe/)[LBuild an issue triage agent with Linear MCPlabels run, state changes wait](https://meandr.com/solutions/linear/)[GHGitHubset rules for each tool](https://meandr.com/solutions/github/)[PHosted Playwright MCP server for web testingruns from the catalog, staging URLs only](https://meandr.com/solutions/playwright/)[#Slack MCP server: let an AI agent post with approvalreads run, posts wait for a person](https://meandr.com/solutions/slack/)[JJira MCP: limit an AI agent to one projectone project key, other projects denied](https://meandr.com/solutions/jira/)[NNotionreads run, page moves wait](https://meandr.com/solutions/notion/)

**MCP SERVERS**

[Connect an MCP server](https://meandr.com/resources/connect-a-server/)[Connect an stdio server](https://meandr.com/resources/connect-a-stdio-server/)[Run a cloud MCP server](https://meandr.com/resources/run-a-cloud-mcp-server/)

**AGENTS**

[Connect an agent](https://meandr.com/resources/connect-an-agent/)[Run a cloud agent](https://meandr.com/resources/run-a-cloud-agent/)

**POLICIES**

[Your first policy](https://meandr.com/resources/first-policy/)[Human approval](https://meandr.com/resources/mcp-human-approval/)[Policy recipes](https://meandr.com/resources/policy-recipes/)

**ACCOUNT**

[Team roles & access](https://meandr.com/resources/team-roles-and-access-control/)

**LEARN MCP**

[What is an MCP gateway?](https://meandr.com/resources/what-is-an-mcp-gateway/)[Gateway vs proxy vs router](https://meandr.com/resources/mcp-gateway-vs-proxy/)[OAuth 2.1 & PKCE for MCP](https://meandr.com/resources/mcp-oauth-pkce/)[MCP 2026-07-28 spec](https://meandr.com/resources/mcp-spec-2026-07-28/)[AWS SigV4 without a proxy](https://meandr.com/resources/sigv4-without-a-proxy/)

**AGENT SECURITY**

[MCP tool poisoning](https://meandr.com/resources/mcp-tool-poisoning/)[OWASP Agentic Top 10](https://meandr.com/resources/owasp-agentic-top-10-mcp/)[Argument-level policy](https://meandr.com/resources/argument-level-policy-vs-allowlists/)[What belongs in an audit log](https://meandr.com/resources/whats-in-an-audit-log/)

[All resources →](https://meandr.com/resources/)

[Pricing](https://meandr.com/pricing/)

[Log In](https://console.meandr.com/)

MCP gateway · Cloud agents · Approvals and audit

# Cloud workspaces for your AI agents and MCP servers

Give each workspace the tools your agents need. Agents keep running 24/7, and you decide what they can do, approve the sensitive parts and see who ran what.

[Try for free](https://meandr.com/pricing/)[Quick tour video](https://www.youtube.com/watch?v=NO3kOHXpxzQ)

SL

Acme’s team

ENworkspaces _0_

Dashboard

System overview

Three layers, built in order. Only the first one needs you right now.

TopologyWaiting for the first connection

**00** ServersConnected

`0` online

Connect or run

Import tools

**00** WorkspacesScoped tool access

`0` ready

Create workspace

Grant access

**00** AgentsWorkspace-bound identities

`0` connected

Connect agent

Resource map

`01 / Infrastructure`_Nothing connected_

`No data`StriperemoteSlackremotePlaywrighthosted

All serversMachines

`02 / Workspaces`_No workspaces_

`No data`Support3 servers · default deny

All workspaces

`03 / Agents`_No agents_

`No data`Claude CodeSupportCodexhosted · Support

All agents

Claude Desktop

Meandrconnected

Add custom connector

Name

Remote MCP server URL

CancelAdd

console.meandr.comOAuth

Claude Desktop wants to use Meandr

As Sam Lee, in Acme’s team.

Read servers, workspaces and logs

Make changes

Answer approvals

CancelAllow

Allow Meandr to access your Stripe account?

Signed in as Sam Lee, account Acme. Meandr keeps the token; the chat never sees it.

CancelAllow access

Connect Slack

Paste a bot token. It goes to Meandr, not to the chat.

Save

Terminal

**$**

Added HTTP MCP server meandr

✓ Signed in · Support workspace connected

MCP servers

## Connect every MCP server, or run your own in the cloud

Paste the URL of any remote MCP server, bring stdio servers from your laptop or office over a tunnel, or have Meandr run one for you. Every server lands in your virtual workspace, ready for your agents.

Servers_8 connected_

- **Jira** Remote`mcp.atlassian.com`connected
- **Linear** Remote`mcp.linear.app`connected
- **Stripe** Remote`mcp.stripe.com`connected
- **Playwright** Hosted`Docker MCP Catalog · EU`running
- **PostgreSQL** Hosted`Docker MCP Catalog · EU`running
- **inventory-mcp** Hosted`your Dockerfile · US`running
- **notes** Tunnel`stdio · sam-macbook`connected
- **app-db** Tunnel`stdio · office-server`connected

Remote **Any Streamable HTTP server**

Paste a URL and pick one of nine sign-in methods. Meandr holds the credential.

Hosted **Run in the cloud, in a few clicks**

From the Docker MCP Catalog, the 20,000+ servers in the Official MCP Registry, or your own Dockerfile.

Tunnel **Your machine, no open ports**

Local files and databases over stdio, from a laptop or any server you run.

[Connect an MCP server →](https://meandr.com/resources/connect-a-server/)[Run a cloud MCP server →](https://meandr.com/resources/run-a-cloud-mcp-server/)

AI agents

## Connect every AI agent, or run your own in the cloud

Connect any agent with a single URL, and it gets instant access to your virtual workspace. Or run your coding agents on Meandr, on machines you order, and pick up the conversation from any device.

- **Claude Code** `$ claude`_Signs in with_Your Claude plan or API key
- **Codex** `$ codex`_Signs in with_Your ChatGPT plan or API key
- **Copilot CLI** `$ copilot`_Signs in with_Your GitHub Copilot plan
- **Cursor CLI** `$ cursor-agent`_Signs in with_Your Cursor account
- **Gemini CLI** `$ gemini`_Signs in with_A Gemini API key or Vertex
- **Goose** `$ goose`_Signs in with_Any provider's key
- **Grok Build** `$ grok`_Signs in with_SuperGrok or an xAI key
- **OpenCode** `$ opencode`_Signs in with_Any provider's key

**One URL, every tool**

Paste your workspace URL into Claude Desktop, ChatGPT, Cursor or any MCP client, and it gets every tool in your virtual workspace. Hosted agents connect in a click.

**Machines you order**

Pick the size, from 512 MB to 8 GB, with shared or dedicated CPU, in the EU or the US. Run several agents on one machine, or give each its own.

**Storage that persists**

Each agent's home is a persistent volume, so its repositories, sign-in and conversations survive restarts and power cycles.

**Your own plan or key**

Each agent signs in with your own subscription or API key, through its own sign-in. Model usage stays between you and its vendor.

[Connect an agent →](https://meandr.com/resources/connect-an-agent/)[Run a cloud agent →](https://meandr.com/resources/run-a-cloud-agent/)

Cloud agents

## Give your agents the tools, and let them work 24/7

Your agents run in the cloud with every tool their workspace allows, around the clock. Check in from your laptop, any browser or your phone: it is one session everywhere, right where you left it.

- 
**Every tool in the workspace**

Connect an agent to a workspace and it gets its servers: GitHub, Linear, a database, a browser. The workspace's policies and approvals apply to every call.

- 
**Works while you sleep**

The agent runs on its own machine for as long as it exists. Close the tab and it keeps going. Nothing stops it for being idle.

- 
**Switch devices, keep the session**

Start on your laptop, check in from your phone's browser, finish anywhere. Every device opens the same session, so closing a tab or switching loses nothing.

- 
**Talk to agents from anywhere**

Use the Meandr console in your web browser. Apps for iOS and Android are coming: answer approvals and check on your agents on the go.

console.meandr.com/agents/hosted/release-bot

SL

**Acme’s team** · Agents › Hosted › **release-bot** _workspaces 1_

HOSTED AGENT

release-bot ready

`Claude Code · FRA · Shared 2 GB · 512 MiB`

RestartStop

**Terminal**

Claude Code runs in it and keeps running after you disconnect. It signs in with /login, with a Claude plan or an Anthropic API key.

Welcome to Claude Code

**\>** Fix the date bug in LIN-482, then open a pull request

**⏺** linear – get\_issue (MCP)(id: "LIN-482")

**⎿** Invoice due dates are off by one day

**⏺** Update(src/invoices.ts)

− const due = new Date( issued );

+ const due = parseIsoDate( issued );

**⏺** Bash(npm test)

**⎿** 214 passed

**⏺** github – create\_pull\_request (MCP)

**⎿** #482 Fix invoice due dates

**\>** Looks good. Merge it once CI passes_from your phone_

**\>**

07:42

**Terminal**

release-bot · the session from your laptop

Welcome to Claude Code

**\>** Fix the date bug in LIN-482, then open a pull request

**⏺** linear – get\_issue (MCP)(id: "LIN-482")

**⎿** Invoice due dates are off by one day

**⏺** Update(src/invoices.ts)

− const due = new Date( issued );

+ const due = parseIsoDate( issued );

**⏺** Bash(npm test)

**⎿** 214 passed

**⏺** github – create\_pull\_request (MCP)

**⎿** #482 Fix invoice due dates

**\>** Looks good. Merge it once CI passes

**\>**

ConnectedHostedMenu

One session on every device

**iOSAndroid_coming soon_**

Control plane

## Put a control plane in front of every call

Every call passes Meandr on its way to a server. Your policies decide it, a person approves it when it matters, and the log keeps it: who called, what they asked for and what happened, live and months later.

Calls today **48,210** 12 agents · 9 servers

Allowed **97.4%** median 318 ms

Held for a person **54** 49 approved · 5 denied

Denied by policy **1,212** before reaching a server

Calls by decision_ **allow**** approve ****deny** _

00:0006:0012:0018:00now

Live log_Support_

- 14:02:19 **release-bot** `github · merge_pull_request`allow
- 14:02:16 **support-bot** `stripe · stripe_api_write`_pending__approved_
- 14:02:13 **qa-agent** `postgres · drop_table`deny
- 14:02:11 **Cursor** `github · create_branch`allow
- 14:02:09 **support-bot** `stripe · stripe_api_read`allow

support-bot · Claude Code

**\>** Refund order 4821, the customer was charged twice

**●** stripe · stripe\_api\_write(PostRefunds, amount: 124000)

MCP server “meandr” requests your input

meandr: approval required — stripe\_api\_write on Stripe

**Comment:** Charged twice, refund approvedOptional note recorded on the audit trail.

**_\*_ OTP:** _K __7__ Q __M__ 2 __R__ V__A_One-time code sent to your email.

AcceptDecline

**●** Refund issued: $1,240.00

ChatGPT

Refund order 4821, the customer was charged twice

Calling meandr · stripe\_api\_write

meandr: approval required — stripe\_api\_write on Stripe_×_

Or approve at: console.meandr.com/actions/01a2…

OTPOne-time code sent to your email.

_K __7__ Q __M__ 2 __R__ V__A_

DenyAccept

Refund issued: $1,240.00

- 
**Approve sensitive actions**

A risky call waits for you. Answer it in the chat, optionally with a one-time passcode (OTP) from the email.

- 
**Hot log**

Every call as it happens: the agent, the tool, the rule that decided and how long it took. Search and filter it for 7 or 30 days, by plan.

- 
**Cold storage**

Older calls move to an archive kept for 30 or 90 days. Download the raw logs on any plan.

- 
**Metrics and history**

Calls, decisions and latency for each workspace and agent, live and by hour or day.

- 
**Insights**

Query months of archived calls, say every denial for one agent since July, and export the result. On Business.

[More about MCP governance →](https://meandr.com/mcp-governance/)[Human approval for tool calls →](https://meandr.com/resources/mcp-human-approval/)

## How Meandr works, step by step

Run servers and agents on Meandr or connect the ones you already have, decide what each agent may call, and see every call it makes.

1. 

01 Meandr MCP

### Set up Meandr from Claude, ChatGPT, Cursor or any MCP client

Add **mcp.meandr.com** to the agent you already use and ask in plain words. It connects MCP servers, creates workspaces and writes policies, and every change shows up in the dashboard.

  - Works with Claude Desktop, Claude Code, ChatGPT, Codex, Cursor and VS Code, or any other AI agent
  - Give your agent read-only or full access
  - OAuth sign-ins and API keys open in your browser, so credentials never pass through the chat

[More information about the Meandr MCP server →](https://meandr.com/resources/connect-an-agent/)

Claude Desktop_Meandr connected_

  - Connect Stripe and Slack **● 2 servers connected** _sign-ins opened in your browser_
  - Run Playwright from the Docker MCP Catalog **● running** _EU · Shared · 512 MB_
  - Put all three in a Support workspace, deny by default **● workspace created** _https://calm-otter.meandr.io_
  - Hold refunds over $300 for my approval **● policy added** _stripe\_api\_write · amount \> 300_
  - What did support-bot do yesterday? **142 calls** _3 denied · 1 approved_

2. 

02 Hosted agents

### Offload your agents to the cloud, then connect and chat like you already do

Run your coding agent on a machine of its own, next to the MCP servers it uses. Connect it to a workspace in one click, and Meandr handles the rest.

  - Its own variables for API keys and tokens, encrypted and never shown again
  - A persistent home: repositories, sign-in and history survive restarts
  - Restart, stop or move it to a bigger machine from the console

Hosted agent · release-bot_ready_

MachineFRA · Shared 2 GB

Memory512 MiB

Homepersistent

Running3 h 12 min
Workspaces

SupportRelease_connected in one click_

Variables
  - `GH_TOKEN`••••••••••••_encrypted_
  - `LINEAR_API_KEY`••••••••••••_encrypted_

Open terminalRestartStop

3. 

03 MCP Hosting

### Give your agents the tools they need with hosted MCP servers

Run a database adapter, a filesystem or a browser for your agents: from the Docker MCP Catalog, the 20,000+ servers in the Official MCP registry, or your own Dockerfile. Pick the machine and we run it, with no CLI and no machine of your own.

  - Several MCP servers can run on one machine: you control what runs there and how many
  - Each server gets its own share of the machine's memory, and each agent that connects its own process
  - EU and US regions, more coming. Shared or Dedicated CPU, arm64 or amd64

[More information about MCP server hosting →](https://meandr.com/mcp-hosting/)

Run a server_Docker MCP Catalog_

Search the directory…

  - **PostgreSQL** Query and update your Postgres database._database_
  - **MongoDB** Query collections and run aggregations._database_
  - **Filesystem** Read, write and search files in a folder._files_
  - **Playwright** Browser automation and testing._browser_

SharedDedicated

arm64amd64

512 MB1 GB2 GB

4. 

04 MCP Gateway

### Connect any server at a URL

Paste the address of any Streamable HTTP server, from Jira and Linear to your own API, and pick how it signs in. Meandr stores the credential and attaches it on the way out, so your agents never hold it.

  - Nine sign-in methods, chosen per server
  - OAuth tokens minted and refreshed for you
  - [AWS SigV4](https://meandr.com/solutions/aws-mcp/) signed server-side, no proxy or sidecar

[More information about authentication methods →](https://meandr.com/mcp-security/#authentication)

Connect a server_Jira_

Server URLhttps://mcp.atlassian.com/v1/mcp

TransportStreamable HTTP

Authentication
  - None
  - Bearer token
  - Custom header
  - Query-string token
  - Basic auth
  - OAuth · client credentials
  - OAuth · authorization code
  - mTLS · client certificate
  - AWS · SigV4

5. 

05 MCP Tunneling

### Connect MCP servers running on your own machine or on any remote server

Give agents your local files or a local database: start the stdio server on a laptop or any server you run with one command. The CLI dials out to Meandr, so no port opens and no credential is stored.

  - Starts your server with the command you already use
  - Reconnects on its own after a restart or a closed laptop
  - Set its policies and limits before it ever connects

[More information about stdio servers →](https://meandr.com/resources/connect-a-stdio-server/)

terminal_sam-macbook_

    # your files, from this laptop$meandr tunnel --id <server-id> -- npx -y @modelcontextprotocol/server-filesystem ~/Projects✓ notes connected · stdio# a database only the office network can reachoffice-server $meandr tunnel --id <server-id> -- npx -y @modelcontextprotocol/server-postgres postgresql://localhost/app✓ app-db connected · stdio reconnected after a restart · 09:41

  - ● **notes** sam-macbook_~/Projects_Connected
  - ● **app-db** office-server_postgresql://localhost/app_Connected

6. 

06 Agents

### Point your agent at Meandr

One workspace URL in the agent's MCP config, and its code stays the same. Claude Desktop signs in over OAuth, so there is no token to paste.

  - Register an agent to get a bearer token for clients that only send a header
  - One key per agent, revoked in one click
  - Agent and server can speak different MCP revisions

[More information about connecting agents →](https://meandr.com/resources/connect-an-agent/)

Connect an agent_Support workspace_

Claude Code

    $claude mcp add --transport http meandr https://calm-otter.meandr.io

Codex

    $codex mcp add meandr --url https://calm-otter.meandr.io

Claude Desktop
  1. Settings → Connectors
  2. Add custom connector
  3. https://calm-otter.meandr.io
  4. Sign in, no token to paste

7. 

07 Security

### Decide what each agent may call

Start with "Let Claude use Stripe" and go as fine as you need. Rules allow, deny or require approval, matched on the agent, the tool, the actual arguments, the caller's IP and headers. They run in priority order, and the first match decides.

  - Flexible policies for every agent, server and tool, in one place
  - A default policy for each workspace decides every call no rule matches
  - Rate limits per second, minute and hour, per workspace, agent or server
  - Team roles: owners and admins reach every workspace, members only the ones you add them to

[More information about access policies →](https://meandr.com/mcp-governance/)[More information about team roles →](https://meandr.com/resources/team-roles-and-access-control/)

Policies_first match wins_

  - 1Let Claude use Stripe_agent Claude Code · server stripe_Allow
  - 2Protect main_github.push\_files · branch = main_Deny
  - 3Refunds over $300_stripe\_api\_write · amount \> 300_Approve
  - 4Office network only_ip not in 10.0.0.0/8_Deny

Default policy_everything no rule matches_Deny

8. 

08 Approvals

### Human-in-the-loop approvals for risky tool calls

A policy can hold a call until a person says yes. Approvers answer from the dashboard or right in the agent's chat on supported clients, and the agent carries on once they do.

  - Approvers notified in the app and by email
  - A one-time code from the approver's email when the policy requires it
  - A call nobody answers is cancelled at its deadline, and the decision is recorded

[More information about approvals →](https://meandr.com/resources/mcp-human-approval/)

Waiting for approval_expires in 14:52_

support-bot wants to refund $1,240.00

ServerStripe

Toolstripe\_api\_write

PolicyRefunds over $300

Notifiedapp · email

Or in the agent's chat

Held for approval. Reply with the code from your email.

_4 __8__ 2 __9__ 1__3_

DenyApprove

9. 

09 Insight

### Watch every call, live and months later

Each call lands in the real-time log with the agent, the tool, the rule that decided and how long it took. Older calls move to cold storage, where you can query and export them.

  - Every decision is recorded, including denials and timeouts
  - Filter the live log by agent, server, tool, status or decision
  - Download raw logs on every plan, and query them on Business

[More information about the audit log →](https://meandr.com/solutions/audit/)

Traffic · Support **Real-time**** Archive ****Insight**

Calls **12,480** last 24 h

Median **318ms** p95 1.9s

Denied **1.4%** 176 calls

Held **23** 21 approved

00:0006:0012:0018:00now

  - 14:02:51 **support-bot** _stripe · stripe\_api\_read_allow212ms
  - 14:02:55 **qa-agent** _playwright · browser\_click_allow1.9s
  - 14:03:02 **Cursor** _notes · write\_file_deny3ms
  - 14:03:14 **support-bot** _stripe · stripe\_api\_write_approveheld

Insight · cold storage

`agent:support-bot action:deny`Jul 1 → Oct 14

14 callsExport CSV

## Built for the data your agents touch

A control plane only works if it stays up and keeps your secrets.

PayloadsMetadata and the request by default. A call's response is kept only when a rule captures it, and everything is encrypted at rest for your plan's retention window.

CredentialsUpstream keys are sealed with AES-256-GCM under a KMS-wrapped data key, one key per environment. Agents hold only a Meandr token, revoked in one click.

In transitTLS 1.2 or newer in transit. Upstream certificates are verified, and the CLI has no insecure mode.

IsolationEvery tenant is isolated along the whole request path, and each of our environments is keyed separately.

AvailabilityRequests enter at the nearest healthy edge location. If a call cannot be recorded, it does not run.

ComplianceBuilt to SOC 2, preparing for the audit. GDPR-aligned: EU SCCs, US processing.

## Questions you'd ask anyway

What is Meandr, exactly?
An MCP Gateway and MCP hosting platform. Your agents connect to one Meandr endpoint, and behind it sit your MCP servers, local or remote, with their credentials kept out of the agent. You can also run MCP servers from the [Docker MCP catalog](https://hub.docker.com/mcp) and the [Official MCP registry](https://registry.modelcontextprotocol.io/) in a few clicks, and run AI agents in the cloud, such as Claude Code, Codex or Gemini CLI, connected to your workspaces. Flexible access policies decide what each agent may call, down to the arguments of a call, and roles decide who on your team manages them. A rule can mark calls for approval, and every call is recorded, making it auditable. [What an MCP gateway is →](https://meandr.com/resources/what-is-an-mcp-gateway/)[Running AI agents on Meandr →](https://meandr.com/resources/run-a-cloud-agent/)
Does it work with the MCP servers I already use?
Yes. An HTTP server connects by its URL, and an stdio server connects through the Meandr CLI on the machine that runs it.
Will my agent connect, or only the well-known clients?
Meandr takes an agent key in a header from any MCP client over HTTP: Claude Code, Cursor, or your own code through the Messages API. It is also an OAuth authorization server, so agents that support OAuth, such as Claude Desktop, connect by signing in: one click, no token to paste.
Which AI agents can I run on Meandr?
Eight coding agents: Claude Code, Codex, Gemini CLI, Copilot CLI, Cursor CLI, Grok Build, Goose and OpenCode. Each runs its vendor's own CLI on a machine you order, you open its terminal in the browser, and it signs in with your own plan or API key. Model usage stays between you and the vendor.
What happens to an agent's files when it restarts?
They stay. Each agent's home is a persistent volume on its machine, so its repositories, sign-in and conversations survive restarts, redeploys and power cycles, and it picks up the last conversation where it left off. Deleting the agent, or the machine it runs on, deletes the volume too.
Can several agents share one machine?
Yes. Each agent needs at least 512 MB of memory, so a 1 GB machine holds one agent with room for an MCP server, and a bigger one holds several. Everything on one machine shares one security boundary, so give an agent a machine of its own when its credentials should stay apart.
What am I actually paying for?
A monthly base for your plan, then the tool calls that reach an upstream and anything beyond what the plan includes. A call your own policies or rate limits block costs nothing. Setting up is free and needs no card. Once you go live, the first 7 days are a free trial. The numbers are on the [pricing page](https://meandr.com/pricing/).
Where does my data actually end up?
On our primary US infrastructure, with Standard Contractual Clauses covering EU, UK, and Swiss transfers. Traffic enters at a regional point close to you before it gets there. Credentials sit encrypted at rest under environment-isolated keys, everything in transit runs TLS 1.2 or newer, and we don't keep upstream payloads in cleartext.
Say an agent's token leaks. Then what?
You revoke it in the dashboard, and Meandr stops accepting it. That's the whole blast radius: your upstream credentials live in our vault, not in your agent, so the leaked token never touched them.
Is any of this open source?
The Meandr CLI is. It is open source under Apache-2.0, so you can read exactly what runs on your machine. The gateway itself is closed-source.

## Host your MCP servers and govern every AI agent call

Meandr runs any MCP server for you, from the Docker MCP catalog, the Official MCP registry or your own container. Agents reach every server through one MCP gateway, with no code changes.

[Try for free](https://meandr.com/pricing/)

Questions? [Email us →](mailto:hello@meandr.com)

Meandr

Cloud workspaces for your AI agents and MCP servers.

- 
- 
- 

#### Platform

- [Overview](https://meandr.com/platform/)
- [MCP Gateway](https://meandr.com/mcp-gateway/)
- [MCP Hub](https://meandr.com/mcp-hub/)
- [MCP Hosting](https://meandr.com/mcp-hosting/)
- [MCP Governance](https://meandr.com/mcp-governance/)
- [MCP Security](https://meandr.com/mcp-security/)
- [AI Agent Security](https://meandr.com/ai-agent-security/)
- [AI Governance](https://meandr.com/ai-governance/)

#### Solutions

- [AWS MCP (SigV4)](https://meandr.com/solutions/aws-mcp/)
- [Prevent dangerous calls](https://meandr.com/solutions/mcp-guardrails/)
- [Monitoring & audit](https://meandr.com/solutions/audit/)
- [Hosted agent workflows](https://meandr.com/solutions/local-mcp-server-remote-agents/)
- [Multi-agent systems](https://meandr.com/solutions/multi-agent-systems/)
- [Stripe](https://meandr.com/solutions/stripe/)
- [Linear](https://meandr.com/solutions/linear/)
- [All solutions →](https://meandr.com/solutions/)

#### Resources

- [Connect an MCP server](https://meandr.com/resources/connect-a-server/)
- [Run a cloud MCP server](https://meandr.com/resources/run-a-cloud-mcp-server/)
- [Connect an agent](https://meandr.com/resources/connect-an-agent/)
- [Run a cloud agent](https://meandr.com/resources/run-a-cloud-agent/)
- [Your first policy](https://meandr.com/resources/first-policy/)
- [Policy recipes](https://meandr.com/resources/policy-recipes/)
- [What is an MCP gateway?](https://meandr.com/resources/what-is-an-mcp-gateway/)
- [All resources →](https://meandr.com/resources/)

#### Company

- [Pricing](https://meandr.com/pricing/)
- [Security](https://meandr.com/security/)
- [Community forum↗](https://community.meandr.com/)

© 2026 Meandr, Inc.169 Madison Ave, STE 66267, New York, NY10016, United States

[Terms of Service](https://meandr.com/terms-of-service/)[Privacy Policy](https://meandr.com/privacy-policy/)[Acceptable Use Policy](https://meandr.com/acceptable-use-policy/)

We use cookies for analytics, ad measurement and remarketing, and our support chat. Decline, and they stay off. [Privacy Policy](https://meandr.com/privacy-policy/)

We use cookies for analytics, ad measurement and remarketing, and our support chat.




## Organization

- [Meandr: Cloud Workspaces for AI Agents & MCP Servers](https://meandr.com/)

## Article

- [Connect an MCP Server over Streamable HTTP](https://meandr.com/resources/connect-a-server/)
- [Write your first policy: allow, deny or ask a person](https://meandr.com/resources/first-policy/)
- [Team roles & access control for MCP workspaces](https://meandr.com/resources/team-roles-and-access-control/)
- [What is an MCP gateway? How it works and who needs one](https://meandr.com/resources/what-is-an-mcp-gateway/)
- [Argument-level policy vs tool allowlists](https://meandr.com/resources/argument-level-policy-vs-allowlists/)
- [SigV4 Signing for MCP: Why a Relay Can’t Do It](https://meandr.com/resources/sigv4-without-a-proxy/)
- [What belongs in an AI agent audit log, call by call](https://meandr.com/resources/whats-in-an-audit-log/)
- [What the MCP 2026-07-28 Spec Changes for Gateways](https://meandr.com/resources/mcp-spec-2026-07-28/)
- [MCP Tool Poisoning: Attacks, Data & Defenses](https://meandr.com/resources/mcp-tool-poisoning/)
- [MCP Gateway vs Proxy vs Router: How to Choose](https://meandr.com/resources/mcp-gateway-vs-proxy/)
- [MCP OAuth 2.1 and PKCE: An Implementer’s Guide](https://meandr.com/resources/mcp-oauth-pkce/)
- [OWASP Agentic Top 10, Mapped to MCP Controls](https://meandr.com/resources/owasp-agentic-top-10-mcp/)
- [MCP Protocol Versions: Mismatched Clients and Servers](https://meandr.com/mcp-protocol-versions/)

## Other Pages

- [AI Agent Control Plane for MCP Servers and Cloud Agents](https://meandr.com/platform/)
- [MCP Hub: One Catalog for Every MCP Server and Tool](https://meandr.com/mcp-hub/)
- [MCP Gateway: One Endpoint for Every MCP Server](https://meandr.com/mcp-gateway/)
- [MCP Governance: Tool Permissions and Approvals](https://meandr.com/mcp-governance/)
- [MCP Security: Secure MCP Servers and Credentials](https://meandr.com/mcp-security/)
- [AI Governance for Agents: Roles, Policies, Audit](https://meandr.com/ai-governance/)
- [Meandr Security: Data Storage, Encryption and SOC 2](https://meandr.com/security/)
- [MCP Elicitation Without Client Support](https://meandr.com/mcp-elicitation/)
- [AWS MCP SigV4: Connect AWS MCP Without a Proxy](https://meandr.com/solutions/aws-mcp/)
- [Solutions: Approvals, Access, Audit, Integrations](https://meandr.com/solutions/)
- [MCP Audit Logs: Find Out What an AI Agent Did](https://meandr.com/solutions/audit/)
- [MCP Gateway for GitHub: One Repository per Agent](https://meandr.com/solutions/github/)
- [Meandr Docs: Set Up and Govern Your MCP Agents](https://meandr.com/resources/)
- [Policy recipes for AI agents: deny, approve, rate-limit](https://meandr.com/resources/policy-recipes/)
- [Pricing: Plans From $4.99/mo, 7-Day Free Trial](https://meandr.com/pricing/)
- [Meandr Terms of Service](https://meandr.com/terms-of-service/)
- [Meandr Privacy Policy](https://meandr.com/privacy-policy/)
- [Meandr Acceptable Use Policy](https://meandr.com/acceptable-use-policy/)
- [Connect an stdio MCP server using a tunnel](https://meandr.com/resources/connect-a-stdio-server/)
- [MCP Hosting: Managed MCP Server Hosting from Docker](https://meandr.com/mcp-hosting/)
- [AI Agent Security: Secure Every Tool Call](https://meandr.com/ai-agent-security/)
- [MCP Guardrails: Block or Approve Dangerous Calls](https://meandr.com/solutions/mcp-guardrails/)
- [Connect a Local MCP Server to ChatGPT & AI Agents](https://meandr.com/solutions/local-mcp-server-remote-agents/)
- [Multi-Agent MCP: Keys, Tools and Limits per Agent](https://meandr.com/solutions/multi-agent-systems/)
- [Stripe MCP Refunds With Human Approval](https://meandr.com/solutions/stripe/)
- [Build an Issue Triage Agent With Linear MCP](https://meandr.com/solutions/linear/)
- [Hosted Playwright MCP Server for Web Testing](https://meandr.com/solutions/playwright/)
- [Slack MCP: Let an AI Agent Post With Approval](https://meandr.com/solutions/slack/)
- [Jira MCP: Limit an AI Agent to One Project](https://meandr.com/solutions/jira/)
- [Notion MCP Server for AI Agents With Approvals](https://meandr.com/solutions/notion/)
- [MCP Human Approval with One-Time Passcodes (OTP)](https://meandr.com/resources/mcp-human-approval/)
- [Run a cloud MCP server from Docker or your own image](https://meandr.com/resources/run-a-cloud-mcp-server/)
- [Connect an AI agent to your MCP gateway](https://meandr.com/resources/connect-an-agent/)
- [Run Claude Code and other AI agents in the cloud, 24/7](https://meandr.com/resources/run-a-cloud-agent/)