/community/guides/agent memory
innernet
docsskillscommunitycompanydeveloperssecurityprivacytermscontact

agent memory · step 6 of 8

your personal agents

Grok Bot, Muse, Hermes, OpenClaw and Instinct, each on one memory. how to connect every one of them, what each can do once it's in, and how to give a personal agent a memory of its own.

/community/guides/agent-memory-personal-agents
6 of 8
step 1 of 83 minmeet agent memoryevery agent you build or use, from a support bot to Grok Bot, Hermes or OpenClaw, can keep a memory of its own. what that memory is, the three ways to make one, and the film of it working. start here.
step 2 of 83 minmake one on the dashboardthe easiest way in, even if you have never wired a memory before. connect an agent in the agents space, paste one block into the ai you code with, and watch it remember.
step 3 of 83 minmake one from the terminalone command in your agent's folder makes the agent and its key, puts the key straight into .env and keeps it out of git. then set what it may do with you, try it, and watch what it keeps, without leaving the terminal.
step 4 of 83 minmake one from any chatany ai connected to innernet can make and run your agents for you. say what you want in plain words, teach the agent one thing and ask it back before any code runs, and let a coding agent wire it in end to end.
step 5 of 84 minwire it into your codean agent's memory is two small calls, one before it answers and one after. here they are in curl, typescript and python, ready to copy, with the one rule for agents that talk to many people, and what comes back.
step 6 of 84 minyour personal agentsGrok Bot, Muse, Hermes, OpenClaw and Instinct, each on one memory. how to connect every one of them, what each can do once it's in, and how to give a personal agent a memory of its own.
step 7 of 83 minwhat it may do with youone choice decides how far an agent goes with you, sealed, asks first or trusted, and four lines say exactly what gets in and out. what it learns about you waits for your yes, and only you can give an agent more of you.
step 8 of 84 minsee everything it keepsan agent's memory is never a black box. its own page shows what it keeps, sorted by kind, every call it made, what its memory did for it, its keys, its sub-agents and the people it serves, and every setting is yours to change.

a personal agent works for you, often for hours, often without you watching: Grok Bot running your Bots, Hermes or OpenClaw on your own machine, Muse in your messages. each one learns how you work, then keeps it to itself, or loses it.

connected to innernet, a personal agent does two things it couldn't before.

  • it reads and writes your one memory, as you. what you told Claude yesterday, Hermes knows today. what Grok Bot learned, ChatGPT can use.
  • it can keep a memory of its own, in the agents space, apart from yours, where you can see everything it keeps and decide what it may do with you.

every connection below uses the same link:

https://innernet.live/api/mcp
every personal agent is a row in connections, with its own steps
every personal agent is a row in connections, with its own steps

Grok Bot

Grok Bot adds an outside memory when you ask it. open the Bot that should remember, and say:

add the custom MCP server innernet at https://innernet.live/api/mcp

then:

  1. 1press Connect on the card it shows.
  2. 2sign in to innernet, and allow it.
  3. 3start a new task. innernet is under Connect apps → Installed.

to have it reach for innernet on its own, put a standing instruction in the Bot's job, in its details. Grok Bot, in every ai has the words to paste.

ask your Bot, press Connect on the card, sign in once
ask your Bot, press Connect on the card, sign in once

Hermes Agent

Hermes (from Nous Research) keeps its outside servers in one file. open ~/.hermes/config.yaml, add this under mcp_servers, and save:

mcp_servers:
  innernet:
    url: "https://innernet.live/api/mcp"
    auth: oauth

then sign in, once, when your browser opens:

hermes mcp login innernet

Hermes Agent, in every ai has its standing instruction, for your AGENTS.md.

one block in config.yaml, one login
one block in config.yaml, one login

OpenClaw

OpenClaw takes two commands, on the machine it runs on:

openclaw mcp set innernet '{"url":"https://innernet.live/api/mcp","transport":"streamable-http","auth":"oauth"}'
openclaw mcp login innernet

the second opens your browser. sign in, and allow it. OpenClaw, in every ai has its standing instruction.

two commands, then a sign-in in the browser
two commands, then a sign-in in the browser

Muse

Muse has no settings page for outside servers. you ask it, in a chat, to build a custom connector, and it does the work on its own.

first, make a key for Muse alone: on your dashboard, open connections → api, name it "muse", keep read and write, and press create key. a key of its own means you can end it any time without touching anything else. then say to Muse:

build a custom connector to the MCP server at https://innernet.live/api/mcp (streamable http). it needs an Authorization header: Bearer, then my innernet key.

when Muse asks for the key, give it in its secure credential prompt, never in the chat. if Muse offers to save the connector as a skill, say yes, so every chat has it.

this is new, and we haven't walked it end to end ourselves yet. if a step reads differently in Muse, tell us and we'll fix this page.

ask Muse for a custom connector, and give the key only in its secure prompt
ask Muse for a custom connector, and give the key only in its secure prompt

Instinct

we haven't found a public way to give Instinct an outside memory yet. if your Instinct account shows a place to add an MCP connection, the link is the same as above, and you sign in to innernet when it asks. when Instinct opens a way in, it becomes one more row in connections, and this page.

and every other agent

Claude Code, Codex, Cursor, Gemini CLI, Windsurf and more connect with one command, npx innernet. ChatGPT, Claude, Perplexity and Le Chat take the link above. connect your ai tools has each one, and every ai, one memory lists them all.

give a personal agent a memory of its own

connected, a personal agent works on your memory, as you. some of what it learns is its own, though: how it runs your inbox, which sources it trusts, what your Bot's followers asked. that belongs in a memory of its own.

ask it, once it's connected:

make yourself an agent in innernet called grok bot, with sub-agents for search, replies and trends. keep it asks first.

the agent and its sub-agents appear in your agents space, each with its own memory. from then on:

  • teach it from any chat: "tell grok bot to never reply to accounts under a week old."
  • ask it back: "what would grok bot read before replying to a new follower?"
  • an agent that runs code (Hermes, OpenClaw, your own) can use its own key and the two calls, so it keeps and recalls on every run. say "give yourself a memory: innernet.live/skills/agent-memory/raw" and it wires itself in.
grok bot's own memory, with the sub-agents it works through
grok bot's own memory, with the sub-agents it works through

next

decide how far each one may go with you: what it may do with you.



every chapter

the fundamentals5 guidesconnecting your ai5 guidesagent memory8 guidesthe dashboard4 guidesfor builders2 guidessharing3 guides

one memory, every ai tool.

innernet is memory your ai tools read live: every project, every decision, every skill, in one place, connected once.

start your innernetevery guide