MaaSLink Documentation

Quick start

After registration, follow the journey for your account type. Enterprise users can configure calls directly; channel users first submit a channel application.

Enterprise users

After registration, you can go directly to the console and configure model access.

01
Register and sign in
Complete account registration and sign in to the MaaSLink console.
02
Create an API key
Create an API key from key management in the console and store the complete key securely.
A complete API key is shown only when it is created. Never commit it to source control or share it unnecessarily.
03
Choose a model and call it
Confirm an available model ID in the model hub or console, then configure your application or tool with the Base URL shown on this page.

Open console

Channel users

After submitting a channel application, you can track its status and provide requested materials. Channel features become available after approval.

01
Register and submit a channel application
Register as a channel user and submit the partnership and qualification details requested on the application page.
02
Check your application status
Use the application status page to see pending, material-required, rejected, or approved status. Add materials there when requested.
03
Use channel features after approval
Once approved, you can associate customers, configure quotas, and view usage. If rejected, follow the page guidance before applying again.

Tool integration

Choose the interface that matches the client: Cursor uses OpenAI Chat Completions, Claude Code uses Anthropic Messages, and Codex uses OpenAI Responses. Follow each tool's own fields for the address, key, and model ID shown on this page.

API and SDK basics

Service address: This page shows the endpoint for the site you are visiting. OpenAI-compatible calls, Cursor, and Codex use a Base URL ending in /v1; Claude Code uses an ANTHROPIC_BASE_URL without /v1.
Authentication: Create an API key in the console and use it only in controlled server-side or tool configuration.
Model ID: Replace MODEL_ID in the examples with a model ID currently available in the console.

Connection values

API Base URL: https://api-intl.rightmaas.com/v1
API key: YOUR_API_KEY
Model: MODEL_ID

Send your first request

curl https://api-intl.rightmaas.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'

OpenAI Chat Completions

Cursor

Configure a Custom Model using Cursor's OpenAI flow. Whether the Custom Model controls are available depends on your Cursor version and plan; if the controls are missing or saving reports a plan restriction, resolve that Cursor-side requirement first.

Override OpenAI Base URL: https://api-intl.rightmaas.com/v1
OpenAI API Key: YOUR_API_KEY
Model: MODEL_ID

Configuration

01
Confirm Custom Model access
Check that your Cursor plan and current version allow Custom Model configuration. If the Models page has no custom-model control or reports a plan error, address that limitation before continuing.
02
Add an OpenAI Custom Model
Open Settings → Models, choose Add Custom Model, and select the OpenAI protocol.
03
Enter the key and Override Base URL
Put the console key in OpenAI API Key, then paste the URL shown on this page into Override OpenAI Base URL. Keep the trailing /v1; do not use the Claude Code Anthropic address here.
04
Paste the exact model ID
Enter the complete model ID shown in the model hub or console. Its casing, hyphens, and version suffix must match exactly.
05
Save and validate
Save the provider, select its model in Cursor chat, and send a non-sensitive test request to confirm it responds.

Anthropic Messages

Claude Code

Claude Code uses the Anthropic Messages protocol. Set ANTHROPIC_BASE_URL to the address shown here without /v1; Claude Code appends /v1/messages when it calls the service.

ANTHROPIC_BASE_URL: https://api-intl.rightmaas.com
ANTHROPIC_AUTH_TOKEN: YOUR_API_KEY
Model: MODEL_ID

Configuration

01
Install or update Claude Code
Install Claude Code using its official method. Do not start its default sign-in flow before configuring the compatible endpoint, so unrelated account credentials are not mixed in.
02
Open settings.json
Use ~/.claude/settings.json on macOS/Linux or .claude/settings.json under your Windows user directory. Create the file if it does not exist.
03
Add the env configuration
Set ANTHROPIC_AUTH_TOKEN, ANTHROPIC_BASE_URL, and the three default-model variables in the env object. The example below uses the address for the site you are visiting; do not append /v1 yourself.
04
Map an available model
Set ANTHROPIC_DEFAULT_OPUS_MODEL, ANTHROPIC_DEFAULT_SONNET_MODEL, and ANTHROPIC_DEFAULT_HAIKU_MODEL to exact model IDs supported in the console. You can begin with the same MODEL_ID for all three.
05
Restart and test
Close the current terminal, open a fresh terminal in your project, run claude, and perform a non-sensitive test task. Confirm use of the API key if prompted.

Configuration example

{
  "env": {
    "ANTHROPIC_AUTH_TOKEN": "YOUR_API_KEY",
    "ANTHROPIC_BASE_URL": "https://api-intl.rightmaas.com",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "MODEL_ID",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "MODEL_ID",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "MODEL_ID"
  }
}

OpenAI Responses

Codex

Codex uses the OpenAI Responses protocol. Its endpoint is /v1/responses, so the provider base_url must retain /v1 and wire_api must be responses.

base_url: https://api-intl.rightmaas.com/v1
experimental_bearer_token: YOUR_API_KEY
Model: MODEL_ID

Configuration

01
Install or update Codex
Install the Codex CLI using its official method and confirm that codex runs in a fresh terminal.
02
Open the configuration file
Use ~/.codex/config.toml on macOS/Linux or C:\Users\<username>\.codex\config.toml on Windows. Create the directory or file if necessary.
03
Add the MaaSLink provider
Put the TOML example below in config.toml, then replace YOUR_API_KEY and MODEL_ID with a key you created in the console and an available model ID.
04
Keep Responses and /v1
Do not change wire_api to chat or completions. The trailing /v1 in base_url corresponds to the /v1/responses route.
05
Restart and validate
Close every running Codex process, start a new project terminal, and run a non-sensitive task. For authentication or model errors, first verify the key, model ID, and TOML syntax.

Configuration example

model_provider = "maaslink"
model = "MODEL_ID"

[model_providers.maaslink]
name = "MaaSLink"
base_url = "https://api-intl.rightmaas.com/v1"
experimental_bearer_token = "YOUR_API_KEY"
wire_api = "responses"

OpenAI Chat Completions

OpenClaw

OpenClaw connects through its openai-completions provider. This guide uses maaslink as a custom provider ID; the provider key and the agents.defaults.model.primary prefix must stay in sync.

baseUrl: https://api-intl.rightmaas.com/v1
apiKey: YOUR_API_KEY
Model: MODEL_ID

Configuration

01
Install OpenClaw
On macOS, Linux, or WSL2 run curl -fsSL https://openclaw.ai/install.sh | bash and follow the installer prompts.
02
Update and confirm the version
Run openclaw update, then openclaw --version to confirm the CLI is installed.
03
Edit the configuration file
Open ~/.openclaw/openclaw.json (create it when needed) and add models.providers.maaslink with this page's /v1 URL, YOUR_API_KEY, and the model id and name.
04
Set the provider and default model
Use the JSON5 example's models.mode: merge and api: openai-completions, then set agents.defaults.model.primary to maaslink/MODEL_ID. maaslink is the custom ID selected for this guide, not an official reserved name; if you rename it, update both the provider key and primary prefix.
05
Validate and troubleshoot
Run openclaw models list, openclaw models set maaslink/MODEL_ID, openclaw models status, openclaw doctor, and openclaw gateway status to check the model, key, and gateway. apiKey may use YOUR_API_KEY or the officially supported ${VAR_NAME} / SecretRef syntax.

Configuration example

{
  models: {
    mode: "merge",
    providers: {
      maaslink: {
        baseUrl: "https://api-intl.rightmaas.com/v1",
        apiKey: "YOUR_API_KEY",
        api: "openai-completions",
        models: [{ id: "MODEL_ID", name: "MODEL_NAME" }],
      },
    },
  },
  agents: {
    defaults: {
      model: { primary: "maaslink/MODEL_ID" },
    },
  },
}

OpenAI Chat Completions

Hermes Agent

Hermes Agent uses a Custom endpoint for a Chat Completions-compatible service. Start with hermes model to configure the provider, authentication, and model, then inspect ~/.hermes/config.yaml when needed.

base_url: https://api-intl.rightmaas.com/v1
api_key: YOUR_API_KEY
Model: MODEL_ID

Configuration

01
Install Hermes Agent
On macOS, Linux, or WSL2 run curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash, then reload your shell.
02
Update and verify the environment
Run hermes update to update the CLI, then hermes doctor to check dependencies and environment.
03
Run the configuration wizard
In a regular terminal run hermes model, choose Custom endpoint, and enter this page's current-site /v1 URL, YOUR_API_KEY, and an available MODEL_ID. hermes model is the complete provider/model wizard: it can add a provider, perform authentication, and save the global default.
04
Understand session switching
Inside an active chat, /model <model> --provider <provider> switches only among configured providers/models; /model <model> --provider <provider> --global also switches the current session and persists the choice in ~/.hermes/config.yaml. /model cannot add a provider, run OAuth, or prompt for an API key; use hermes model for those actions.
05
Inspect the manual configuration
Alternatively edit ~/.hermes/config.yaml and set model.default, provider: custom, base_url, api_key, and api_mode: chat_completions under model. Check it with hermes config get model --json, hermes status, and hermes doctor. api_key may use YOUR_API_KEY or ${VAR_NAME} / ${env:VAR_NAME} references.

Configuration example

model:
  default: MODEL_ID
  provider: custom
  base_url: https://api-intl.rightmaas.com/v1
  api_key: YOUR_API_KEY
  api_mode: chat_completions

Frequently asked questions

These answers help you quickly resolve common setup and account questions.

How do I obtain an API key?

Sign in to the console and create an API key in key management. Use the key only in controlled environments; never put it in client-side code, public repositories, or screenshots.

Which API Base URL should I use?

Use the address shown on this page for the site you are visiting. OpenAI-compatible calls, Cursor, OpenClaw, and Hermes Agent use the URL ending in /v1; Codex uses OpenAI Responses and must also retain /v1; Claude Code's ANTHROPIC_BASE_URL omits /v1 because the client appends /v1/messages. Do not mix credentials or addresses between sites.

Where can I find a model ID?

Use a model ID currently available in the model hub or console for this site. Availability can change, so do not use MODEL_ID from the example directly in production.

How are requests authenticated?

For server-side requests, send the API key as Authorization: Bearer YOUR_API_KEY. When a tool uses its own authentication fields, map them according to that tool's official documentation.

What can I do while a channel application is reviewed?

You can view application status and add requested materials. Customer association, quota configuration, and usage operations become available only after approval.

Where can I review usage and costs?

Use the console to review account balance, call usage, and billing information. Channel quota and customer-level data follow the channel workflow available after approval.

Terms and agreements

Review the terms and agreements that apply to your use of MaaSLink services.

User Agreement
Review the rules and responsibilities for using this service.
View agreement
Privacy Policy
Learn how your personal information is collected and used.
View agreement

MaaSLink Documentation

Quick start

Enterprise users

Register and sign in

Create an API key

Choose a model and call it

Channel users

Register and submit a channel application

Check your application status

Use channel features after approval

Tool integration

API and SDK basics

Connection values

Send your first request

Configuration

Confirm Custom Model access

Add an OpenAI Custom Model

Enter the key and Override Base URL

Paste the exact model ID

Save and validate

Configuration

Install or update Claude Code

Open settings.json

Add the env configuration

Map an available model

Restart and test

Configuration example

Configuration

Install or update Codex

Open the configuration file

Add the MaaSLink provider

Keep Responses and /v1

Restart and validate

Configuration example

Configuration

Install OpenClaw

Update and confirm the version

Edit the configuration file

Set the provider and default model

Validate and troubleshoot

Configuration example

Configuration

Install Hermes Agent

Update and verify the environment

Run the configuration wizard

Understand session switching

Inspect the manual configuration

Configuration example

Frequently asked questions

Terms and agreements

User Agreement

Privacy Policy