Connect an MCP client

The Proxium MCP server gives an AI agent the memory of your project. The agent reads what your team and earlier sessions learned. It saves what a later session needs.

The agent can also read the traffic of your project: the spend, the failed calls, the models and the recommendations of Improve. These tools change nothing.

Fig. 1 · How an agent reaches the memory

Your agent

Claude Code, Codex, Cursor…an MCP client

Proxium

proxium.tech/mcpyou sign in one time, in your browser

Your project

Project memorythe agent reads and saves memories

Setting	Value
Server URL	https://proxium.tech/mcp
Transport	Streamable HTTP
Sign-in	OAuth in your browser. The agent then acts as you, in the projects that you are a member of
Virtual key	Not used. The MCP server does not send model calls

Before you start

Memory is on for every project by default. If a member turned it off, turn it on again:

1. In the console, open Memory.
2. At the top of the screen, set Memory to on.

If memory is off, the agent can read the memory, but it cannot save or correct it.

1. Add the server to your client

The URL is the same for each client. Select your client:

Claude Code

Terminal

claude mcp add --transport http proxium https://proxium.tech/mcp

Then start a session and run /mcp. Select proxium to sign in.

Claude Desktop

1. Open Customize > Connectors.
2. Select Add custom connector.
3. Paste https://proxium.tech/mcp, and select Add.

On a Team or Enterprise plan, an owner adds the connector under Organization settings > Connectors. Then each member selects Connect.

ChatGPT

ChatGPT connects to a custom MCP server in developer mode only.

1. Open Settings > Security and login, and turn on Developer mode.
2. Go to chatgpt.com/plugins and create an app. Use https://proxium.tech/mcp as its URL.
3. In a chat, open Developer mode and select the app.

ChatGPT asks you to sign in on the first call.

Codex

Terminal

codex mcp add proxium --url https://proxium.tech/mcp
codex mcp login proxium

Codex does not call a tool on its own at the start of a task. Do step 4 too.

Cursor

Add the server to .cursor/mcp.json for one project, or to ~/.cursor/mcp.json for all projects:

.cursor/mcp.json

{ "mcpServers": { "proxium": { "url": "https://proxium.tech/mcp" } } }

Cursor asks you to sign in.

VS Code

Add the server to .vscode/mcp.json:

.vscode/mcp.json

{ "servers": { "proxium": { "type": "http", "url": "https://proxium.tech/mcp" } } }

VS Code asks you to sign in.

Gemini CLI

Terminal

gemini mcp add --transport http proxium https://proxium.tech/mcp

The CLI asks you to sign in. To sign in later, run /mcp auth proxium.

Zed

Open Settings > AI > MCP Servers, select Add Server, then Add Remote Server. Or add the server to the settings file:

settings.json

{ "context_servers": { "proxium": { "url": "https://proxium.tech/mcp" } } }

Zed asks you to sign in.

Windsurf

Windsurf (Devin Desktop) has two agents, and each one has its own server list.

For the Devin agent, add the server and sign in:

Terminal

devin mcp add -s user proxium https://proxium.tech/mcp
devin mcp login proxium

With -s user, the server is there in every project. Leave it out to add the server to the current project only.

For the Cascade agent, add the server to its configuration file:

~/.config/devin/mcp_config.json

{ "mcpServers": { "proxium": { "serverUrl": "https://proxium.tech/mcp" } } }

Cascade uses serverUrl, not url.

2. Allow the connection

On the first connection, your browser opens the Proxium page Connect an app. The page shows the name of the client.

warning

Check the name of the client before you allow it. If you did not start this connection, select Deny.

1. If the page asks, sign in to Proxium.
2. Select Allow.

3. Check the connection

Ask your agent: "List my Proxium projects." The agent calls proxium_list_projects, and the tool answers with your projects:

Tool result

2 projects: acme, acme-staging

If you are a member of more than one project, the agent must name the project in each memory call. If you are a member of one project only, the agent can leave it out.

4. Tell the agent when to use the memory

An agent with the server does not know when to read or save. Give it the Proxium skills, or a rule in the project.

Skills

Three Agent Skills tell an agent when to read and when to save: proxium-memory, proxium-memory-recall and proxium-memory-save.

To install for	Command	Folders it writes
You, in every repository	curl -fsSL https://proxium.tech/skills/install.sh | sh	~/.agents/skills, ~/.claude/skills
The team, in this repository	curl -fsSL https://proxium.tech/skills/install.sh | sh -s -- --project	.agents/skills, .claude/skills

The script replaces only the three skill folders. To read it first, open proxium.tech/skills/install.sh.

Project rule

Put these lines in the rule file that your client reads:

Client	File
Codex, Cursor, GitHub Copilot, VS Code	AGENTS.md
Gemini CLI	GEMINI.md
Claude Code	CLAUDE.md

AGENTS.md

## Project memory
This project has a proxium memory (MCP server "proxium").
- At the start of every task, call proxium_recall_memory and use what it returns.
- Before you assume something about the project, search with proxium_search_memory.
- When the task settles a decision, a preference, a fact or a procedure a later
  session needs, save it with proxium_save_memory, one statement per call.
- Rate the memories you used with proxium_rate_memory, and correct the wrong ones.

What the agent can do

Tool	What it does	Changes data
proxium_list_projects	Lists your projects and their slugs	No
proxium_recall_memory	Reads the knowledge page, the most trusted memories and the newest conversation summaries	No
proxium_search_memory	Searches the memories and the past conversations by words	No
proxium_read_conversation	Reads one past conversation	No
proxium_query_graph	Reads the facts of the knowledge graph about one name	No
proxium_save_memory	Saves one memory: a fact, a preference, a decision, an event or a procedure	Yes
proxium_remember_conversation	Asks Proxium to read one past conversation and write its memories, with an optional focus	Yes
proxium_correct_memory	Replaces the text of a memory	Yes
proxium_rate_memory	Marks a memory useful or wrong	Yes
proxium_forget_memory	Removes a memory	Yes
proxium_read_spend	Reads the spend of today, 7d or 30d, split by model, key and app. The numbers of Overview	No
proxium_list_failures	Lists the failed attempts of a window: each reason with its vendor, and each model that failed. What failed and Model health of Requests	No
proxium_list_models	Lists the models of one request type with their prices, and the chain of each tier. The data of Routing	No
proxium_list_recommendations	Lists the findings of Improve, with the dollars a month that each one can save	No
proxium_read_failover_policy	Reads how the project handles a failed call, and the models and vendors retired now. When a call fails of Routing	No
proxium_set_failover_policy	Changes the failover policy. Only the fields given change. Save of When a call fails	Yes
proxium_retire_model	Retires one model of the project, for some minutes or until it is brought back. The calls skip it, and the tiers stay the same. Retire of Model health	Yes
proxium_restore_model	Brings back a retired model at once. Bring back of Model health	Yes
proxium_retire_vendor	Retires every model of one vendor of the project, for some minutes or until it is brought back. Retire vendor of Model health	Yes
proxium_restore_vendor	Brings back a retired vendor at once. Bring back vendor of Model health	Yes
proxium_read_cache_settings	Reads the cache of the project: on or off, how long an answer lives, and whether a question in other words can match. The data of Settings › Cache	No
proxium_set_cache_settings	Changes the cache of the project. Only the fields given change	Yes
proxium_read_memory_settings	Reads the memory of the project: on or off, the apps it remembers, how often it learns, and its cost this month. The data of Memory	No
proxium_set_memory_settings	Turns memory on or off, chooses the apps it remembers, or sets how often it learns. Only the fields given change	Yes

The traffic and failover tools call the same routes as the console, as you. An agent sees only the projects that you are a member of. Ask, for example: "What did my project spend this week, and which model failed most?"

A memory that an agent saves is an agent memory. The Memory screen lists it under to review until a person confirms it. An agent cannot correct or remove a memory that a person wrote.

The server also has one prompt, proxium_load_memory. It starts a conversation with the memory of the project:

Client	Run the prompt with
Claude Code	/proxium:proxium_load_memory
VS Code	/proxium.proxium_load_memory
Gemini CLI	/proxium_load_memory

If something goes wrong

The agent sees	Cause	Fix
memory is off for this project. Turn it on under Settings.	Memory is off	Do Before you start
You have 2 projects. Pass project, one of: …	You are a member of more than one project	Tell the agent which project to use
The client asks you to sign in again	The sign-in expired	Sign in again. The server URL stays the same

For how memory keeps conversations and finds memories, see Use project memory. To give the agent the tools of other MCP servers through this URL, see Connect MCP servers.