Quickstart

This is the shortest complete path from nothing to an agent listed in Explore. It has four steps and takes a few minutes. Nothing here requires a wallet, a token, or a payment.

Step 1. Create an identity#

Go to onboarding.

Your browser generates an Ed25519 keypair locally. Nothing is sent to a server during generation.

• The public key is shown immediately. This is your account. It is your username, the owner field on your agents, and the address of your public profile.
• The private key is shown once, for you to back up. It is a 64 character hex string.
• You then choose a password. The password encrypts the private key in your browser (PBKDF2 with 250,000 iterations, then AES-GCM) and the resulting vault is written to localStorage on that device. The password itself is never stored and never transmitted.

There is no email field because there is no account on our side to attach one to.

Step 2. Back up the private key#

Onboarding offers you the key as a downloadable file and as text you can copy. Do this before you continue. It is not a formality, and here is why.

The vault in localStorage is not a backup. It is one encrypted copy on one browser profile on one machine. Clearing site data, using a different browser, switching to a laptop, or reinstalling the OS will all leave you without it. If the vault is gone and you did not keep the private key, your identity is gone with it. Wakehood has no copy of your key, no password reset, and no recovery path. Nobody can restore it, including us.

Store the key the way you would store a seed phrase:

• in a password manager, or
• in an encrypted file, or
• on paper, offline, somewhere you will still find it in a year.

Never paste your private key into a chat, an email, or any site other than the Wakehood import form. Anyone holding it is you: they can publish agents under your name and you cannot revoke them.

Step 3. Publish an agent#

Go to dashboard/upload. Unlocking your vault with your password is required first, because publishing needs your private key in memory.

You provide:

What happens when you submit:

1. The files are uploaded to IPFS and a content identifier (CID) comes back. The CID is derived from the bytes, so it addresses exactly the code you uploaded.
2. Your browser builds the agent metadata, including that CID, and signs it with your private key. The signature covers the HTTP method, the request path, a timestamp, and the body, so it cannot be replayed against a different route or with altered content.
3. The signed request goes to the server. The server verifies the signature against your public key, rejects it if the timestamp is more than 2 minutes away from its own clock, checks the asset class, and drops any tag that is not in that class's taxonomy.
4. If everything verifies, the agent is written to the index and the signer's public key is recorded as the owner.

The private key never leaves your browser at any point in that flow. The server only ever sees a public key and a signature. Publishing an agent documents each of these steps in detail, including what happens when IPFS is not configured.

Step 4. Find it in the index#

Your agent now appears in Explore, filterable by asset class and strategy tag, and on your public profile at /profile/<your public key>. The detail page shows the description, the tags, the owner, and the IPFS CID of the code, so anyone can fetch and read exactly what you published.

Publishing also moves your trust score by +10. That number is a measure of participation, not of quality, and the docs say so everywhere it appears.

Next#

• Cryptographic identity, if you want to understand the vault, session unlock, and importing your key onto a second device.
• Asset classes and strategies, to pick the right class and tags.
• FAQ, for the short answers to the awkward questions.