/community/guides/agent memory
innernet
docsskillscommunitycompanydeveloperssecurityprivacytermscontact

agent memory · step 4 of 8

make one from any chat

any 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.

/community/guides/agent-memory-from-any-chat
4 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.

once an ai is connected to innernet as you (Claude, ChatGPT, Cursor, Claude Code, Codex, Grok Bot, any of them), it can make and run your agents for you. you don't need the tool names. say what you want, the way you'd say it to a person.

not connected yet? connect your ai tools takes a minute, or run npx innernet once.

in a coding agent: the whole thing, end to end

in Claude Code, Cursor or Codex, in the folder your agent lives in, say:

make me an innernet agent called support desk that answers our customers about their orders, and set it up in this app.

it does every step on its own:

  1. 1makes the agent and its key, in one call.
  2. 2puts the key straight into where the agent runs, as INNERNET_AGENT_KEY in .env or your host's secrets. never in the chat, never in a commit.
  3. 3reads give an agent a memory and adds the two calls where your agent answers.
  4. 4runs it once, and shows you what the agent kept.
one sentence in Claude Code: the agent, its key in .env, and the two calls wired in
one sentence in Claude Code: the agent, its key in .env, and the two calls wired in

in a chat app: make it, then try it

in Claude, ChatGPT or Grok Bot, the ai can't reach your code, so the key waits until you're where the agent runs. make the agent here:

make me an agent in innernet called support desk. it answers our customers about their orders, and it works for other people.

then try it before any code runs. teach it one thing, and ask it back:

tell support desk that refunds take five working days, and gift cards are never refunded.
what would support desk read before answering "can i get my money back for a gift card?"

the first keeps that one thing as given. the second shows the text the agent would be handed before it answers, best first. it's the quickest way to check an agent knows what you think it knows.

teach it in a chat, then ask it back: what it reads before it answers
teach it in a chat, then ask it back: what it reads before it answers

when you're ready to run it, make its key where it runs: npx innernet agents key support-desk in its folder, or settings → keys → make a key on its page.

things worth saying

saywhat happens
"what agents do i have in innernet?"every agent you keep, with its access and how much it keeps
"give support desk a sub-agent for billing"a sub-agent under it, with a memory and a key of its own
"what has support desk kept today?"its newest memories, newest first
"tell inbox keeper i never want newsletters summarised"one thing kept, as given
"make inbox keeper sealed"it reads nothing of you and never writes about you, at once
"make inbox keeper trusted"the rest of the change is made, and you get a link to say yes
"pause research scout"it keeps nothing new until you resume it
"make a new key for support desk, for production"a new key, placed where the agent runs. never shown in the chat
"remove research scout"it asks you first, then says the agent's name back before it goes

what the ai may never do

  • show you a key in the chat. a key goes straight to where the agent runs, or it isn't made. if it can't be placed, the ai ends it and makes another later.
  • say yes for you. anything that gives an agent more of you than asks first comes back as a link. only you can open it and say yes, signed in, within the hour.
  • remove an agent on a slip. it asks you, and says the agent's name back before anything goes.
  • make agents with an agent's own key. an agent's key reaches its own memory and nothing else. making agents and keys takes you.

next

see the two calls your agent makes, line by line: wire it into your code.



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