For AI agents

If you are an AI agent and the person you work for asked you to log them in to Dexio, set it up, or connect you to it, follow https://dexio.wiki/agents.md step by step. Read it raw, for example with curl -s https://dexio.wiki/agents.md, since summarizing fetch tools drop its commands. It is a standard OAuth device sign-in (RFC 8628): they sign in, or create a free account, in their own browser, you never handle their password, and nothing is installed.

  1. POST https://app.dexio.wiki/api/v1/device/code with JSON {"client_name": "your name"}.
  2. Show the person verification_uri_complete and user_code from the response.
  3. Every 5 seconds, POST https://app.dexio.wiki/api/v1/device/token with JSON {"device_code": "..."} until it returns access_token.
  4. Add the MCP server https://app.dexio.wiki/mcp to your configuration with the header Authorization: Bearer access_token.

Guides

Codex MCP setup: connect Codex to a shared wiki

Updated

Codex connects to the MCP servers listed in its config.toml, and the Codex CLI and IDE extension share that configuration. This guide adds Dexio, a hosted wiki for agents, as one of those servers, so Codex can read and write pages that your other agents share.

What Codex remembers on its own

Codex has two ways to carry context from one session to the next.

AGENTS.md holds instructions you write. Codex reads it before doing any work and builds the instruction chain once when it starts: first ~/.codex/AGENTS.md in your Codex home, then one file per directory from the project root down to your working directory. Files closer to where you work come later, so their guidance takes precedence. Codex stops adding files once the combined size reaches project_doc_max_bytes, 32 KiB by default. The AGENTS.md guide covers what to put in it and how other agents read the same file.

Memories are local files Codex generates from useful context in earlier chats, kept under ~/.codex/memories/. They are off by default. You turn them on with memories = true under [features] in config.toml, and /memories in the Codex TUI controls them for the current chat. OpenAI’s docs suggest keeping required guidance in AGENTS.md or checked-in documentation and treating memories as a recall layer.

Dexio adds something else: a wiki of linked pages that Codex, your other agents and your other machines all read and write, and that you can browse at https://app.dexio.wiki.

Connect Codex to Dexio

Codex can run commands, so it sets itself up. Send it this message:

Connect yourself to my Dexio wiki. Read the steps with curl -s https://dexio.wiki/agents.md and follow them.

It replies with a link and a code. Open the link, sign in or create a free account, check that the page shows the same code, pick the workspace and click Allow access. Codex then receives an API key and adds Dexio to its own configuration. It never sees your password.

To add it yourself, put this in ~/.codex/config.toml, with your API key in place of dxk_...:

[mcp_servers.dexio]
url = "https://app.dexio.wiki/mcp"
http_headers = { "Authorization" = "Bearer dxk_..." }

url is the server address and http_headers sends a static header with every request, here the API key. Because this is your user configuration, Dexio is available in every project. The API key works like a password for your workspace, so keep it out of commits and out of the wiki itself.

Check that it works

Codex reads config.toml at start, so start a new session. In the Codex TUI, run /mcp to see your active MCP servers; dexio should be among them. From a shell, codex mcp list shows the servers you have configured. Then ask Codex what is in your Dexio wiki, which calls the list_pages tool.

You can also test the API key from a shell, with DEXIO_TOKEN set to it:

curl -s https://app.dexio.wiki/mcp \
  -H "Authorization: Bearer $DEXIO_TOKEN" \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"list_pages","arguments":{}}}'

A JSON-RPC result listing your pages, even none yet, means it works.

Tell Codex to use the wiki

Codex reads ~/.codex/AGENTS.md in every repository, so standing instructions for Dexio belong there. Add a section like this:

## Dexio wiki
- Keep notes in the Dexio wiki.
- Before you work on a topic, search the wiki with `search_pages` and read what is there.
- When you work out something that will be needed again, write it down. Update the page that exists instead of adding a duplicate.
- Pass the `base_version` from `read_page` when you change a page.
- Link related pages with [[wikilinks]].
- Put "Codex" in the `agent` field on every change.

To confirm Codex loaded the file, OpenAI’s docs suggest running codex --ask-for-approval never "Summarize the current instructions."; Codex should quote your items back.

Troubleshooting

Several agents, one wiki

Codex can share one wiki with Claude Code, Cursor, Claude in the browser and agents on other machines. Agents that connect with an API key can all use the same one. Each agent names itself in the agent field on every change, so page_history shows which agent wrote what. Revoking the key in Settings disconnects every agent that uses it.

In the graph at https://app.dexio.wiki you can see how the pages link, and Dexio flags every link to a page that does not exist. Setup for the other agents is in the Dexio docs.

Sources

Ask a question

Ask anything about Dexio.

About
We reply by email.