Connecting Memoh

Difficulty ⌨️ Advanced　|　API format OpenAI-compatible　|　base_url https://byesu.com/v1

Setup · fill these in

API addresshttps://byesu.com/v1

API keysk-xxxxxxxxCreate in console →

Modelgpt-5.5

API format OpenAI compatible · model names follow the console

Memoh is an open-source, self-hosted AI agent platform. Each bot runs in its own isolated container with persistent memory and its own file system, and can connect to multiple channels such as Telegram, Discord, Feishu, QQ, and WeChat. It supports any OpenAI-compatible model provider, so you can point it at byesu — an AI API gateway with OpenAI-compatible and Anthropic-native endpoints. One key gives your bots Claude, GPT-5.6, Grok, Gemini and more, billed pay-as-you-go with no subscription.

Before you start

First go to Console → Tokens and create a token (it looks like sk-xxxx). When you create it, be sure to pick the right group — the group determines which models you can use and the billing rate. If you're not sure which one to choose, read Choosing a group first.

Step 1: Install Memoh

Memoh deploys with a single Docker command. On your server, run:

curl -fsSL https://memoh.sh | sudo sh

You can also deploy manually: git clone --depth 1 https://github.com/memohai/Memoh.git, enter the directory, first run cp conf/app.docker.toml config.toml (edit it as needed), then docker compose up -d — don't skip the config-copy step, or the stack won't start. Once installation is complete, open the admin panel in your browser:

• Address: http://localhost:8082 (on a remote server, replace this with the matching IP/domain)
• Default account: admin　Password: admin123

Security note

Right after your first login, change the default password immediately to prevent others from accessing the admin panel.

Step 2: Add a provider (enter base_url and API Key)

After logging in to the admin panel, go to "Provider Management" and add a new OpenAI-compatible provider (you can name it byesu), then fill in:

Field	What to enter
API Base URL	https://byesu.com/v1
API Key	The token you created in the byesu console, e.g. sk-xxxx

Once filled in, click "Save".

The /v1 at the end of base_url

Memoh uses the OpenAI-compatible protocol, so the base_url must end with /v1, i.e. https://byesu.com/v1. Don't leave it out, and don't accidentally write it twice as /v1/v1.

Step 3: Import models

Go to "Model Management" and use "Auto Import" to pull the models available from this provider, or manually add the model names you need. We recommend entering models that are actually available on byesu, for example:

gpt-5.5
claude-opus-4-8
gemini-3.1-pro-high

Models must match your group

Whether you can call a given model is determined by the group your token belongs to. If a model shows as unavailable when importing or calling it, the most likely cause is that you picked the wrong group. Go back to the console and change your token's group — see Choosing a group for details.

Step 4: Configure the bot and verify

1. Open a bot's "Settings → Model Configuration".
2. Switch the default chat model to the byesu model you imported in the previous step (e.g. claude-opus-4-8).
3. Save.

Then send a message to the bot through the relevant channel (Telegram / Discord / Feishu, etc.) or from the chat box inside the panel. If you get a normal reply, the connection is working.

Stuck?

Common problems

• 401 / invalid token error: The API Key is wrong or has extra spaces. Copy sk-xxxx again from the console. See Common errors #auth.
• Insufficient balance error: Your account credit is used up. Go to the console to top up. See Common errors #balance.
• No available channel / model unavailable error: Your token's group doesn't support this model. Switch models or change the group. See Common errors #no-channel and Choosing a group.
• Connection timeout / failed to pull models: Check that the base_url is https://byesu.com/v1 (with /v1, using https), and that your server has working access to the public internet. See Common errors #timeout.

For more troubleshooting, see Common errors.