Roo Code is a Cline fork for VS Code, configured the same way. A third-party key — no restrictions.
- Install the Roo Code extension from the VS Code Marketplace.
- Settings → API Provider → Anthropic.
- Check Use custom base URL →
https://api.guardrelay.ai. - API Key —
gd-…from the bot. - Model —
claude-sonnet-5-5orclaude-opus-5.
Kimi works with either provider. For Anthropic, set the Custom Base URL to https://api.guardrelay.ai; for an OpenAI-compatible one, set Base URL to https://api.guardrelay.ai/v1. The key is the same gd-… and the balance is shared, and Model takes one of the four Kimi IDs.
Both addresses work: https://api.guardrelay.ai on the Anthropic provider and https://api.guardrelay.ai/v1 on an OpenAI-compatible one. Put the exact id in Model, for example cursor-grok-4.6-medium or composer-2.5-fast.
This guide applies to an existing Roo Code 3.54 installation. For a new installation, prefer the maintained Cline. Open a separate test folder and the Roo Code panel in VS Code.
Get a universal gd-… key from the direct API link in @guardapi and choose an API package. Do not put the key in a project, repository, task prompt, or screenshot. Balance and usage are available in your Guard account.
- Select the Roo Code gear and choose API Provider → OpenAI Compatible. Do not choose Native Responses.
- Set Base URL to
https://api.guardrelay.ai/glm/v1, API Key togd-…, and Model toglm-5.3. - Keep Enable streaming on. Disable Enable R1 model parameters, Use Azure, Enable Reasoning Effort, and custom temperature; leave Custom Headers empty.
- If Include max output tokens is enabled, set Max Output Tokens to
4096as a bounded client example. Do not treat Context Window Size as the GLM limit or publish it as a model property.
| Roo Code field | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://api.guardrelay.ai/glm/v1 |
| API Key | gd-… |
| Model | glm-5.3 |
| Max Output Tokens | 4096 |
| Final request | POST https://api.guardrelay.ai/glm/v1/chat/completions |
- Open a new task and send:
Read README.md and reply with exactly GUARD_GLM_OK. Change nothing. - Approve only reading. Then, in the same task, ask for the line without reopening the file.
- If the first command reports a missing
rgor cannot find files, this is a Roo client dependency issue: reinstall/update it or use Cline.
| Model ID | When to use it |
|---|---|
glm-5.3 | Default model; text and coding tasks. |
glm-5.3-flash | Fast model; at API level it accepts PNG and PDF through all three protocols, and MP4 through Chat Completions only. |
- With the stated dependency qualification, streaming Chat, local tools, and same-task continuation are confirmed.
- The client itself executes file, edit, and terminal tools. This profile does not connect Z.AI tools or MCP automatically.
- The Chat API permits MP4 with Flash and PNG/PDF at API level; attachments through the Roo UI were not checked.
- The profile uses OpenAI-compatible Chat Completions. The GLM server prompt is off by default for ordinary API access; client setup does not enable it.
- Native TXT, DOCX, and XLSX are outside the confirmed GLM contract. Use PDF for documents, or PNG with
glm-5.3-flashfor images.
| Symptom | What to check |
|---|---|
401 invalid_api_key | The client received the complete active gd-… key; the variable is set in the same process that starts the client. |
404 model_not_found | The model ID is exactly glm-5.3 or glm-5.3-flash; old GLM IDs, aliases, Turbo, and FlashX are unsupported. |
404 or incorrect path | Base URL is https://api.guardrelay.ai/glm/v1. The client appends /chat/completions; do not add it twice. |
400 invalid_request | The documented API format is selected; reasoning, temperature, MCP, or other unconfirmed fields were not copied from an older profile. |
402 insufficient_balance | Check available balance and unfinished reservations in your Guard account. |
| The model is replaced or the request uses another route | Choose OpenAI Compatible, exact glm-5.3, and disable Native Responses, R1, and Reasoning Effort. |
For a new installation, prefer the maintained Cline. Open a separate test folder and the Roo Code panel in VS Code.
Get a universal gd-… key from the direct API link in @guardapi and choose an API package. Do not put the key in a project, repository, task prompt, or screenshot. Balance and usage are available in your Guard account.
- Select the Roo Code gear and choose API Provider → OpenAI Compatible. Do not choose Native Responses.
- Set Base URL to
https://api.guardrelay.ai/v1, API Key togd-…, and Model togemini-4-argon. - Keep Enable streaming on. Disable Enable R1 model parameters, Use Azure, Enable Reasoning Effort, and custom temperature; leave Custom Headers empty.
- If Include max output tokens is enabled, set Max Output Tokens to
4096as a bounded client example.
| Roo Code field | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://api.guardrelay.ai/v1 |
| API Key | gd-… |
| Model | gemini-4-argon |
| Max Output Tokens | 4096 |
| Final request | POST https://api.guardrelay.ai/v1/chat/completions |
- Open a new task and send:
Read README.md and reply with exactly GUARD_GEMINI_OK. Change nothing. - Approve only reading. Then, in the same task, ask for the line without reopening the file.
| Model ID | Input and window |
|---|---|
gemini-4-argon | Text, images, PDF; 1,048,576-token window, shared by input and answer. |
gemini-3.8-flash | Text, images, PDF; 1,048,576-token window for input. |
- Roo Code itself executes file and terminal tools; it sends the task history to Gemini with every request.
- The Guard catalog takes text, images, and PDF as input and returns text. How to send a file or an image depends on the client's interface; this guide does not describe attachments.
- The profile uses OpenAI-compatible Chat Completions. The server prompt is a separate setting of the Gemini family: it applies only when the key has the entitlement and the option is on; client setup does not enable it.
| Symptom | What to check |
|---|---|
401 invalid_api_key | The client received the complete active gd-… key; the variable is set in the same process that starts the client. |
404 model_not_found | The model ID is an exact ID from the table above or the short name gemini; other spellings are unsupported. |
404 or incorrect path | Base URL is https://api.guardrelay.ai/v1. The client appends /chat/completions; do not add it twice. |
400 invalid_request | The documented API format is selected; extra settings from an older profile were not copied: reasoning, temperature, MCP, and others. |
402 insufficient_credit | Check available balance and unfinished reservations in your Guard account. |
| The model is replaced or the request uses another route | Choose OpenAI Compatible, exact gemini-4-argon, and disable Native Responses, R1, and Reasoning Effort. |
For a new installation, prefer the maintained Cline. Open a separate test folder and the Roo Code panel in VS Code.
Get a universal gd-… key from the direct API link in @guardapi and choose an API package. Do not put the key in a project, repository, task prompt, or screenshot. Balance and usage are available in your Guard account.
- Select the Roo Code gear and choose API Provider → OpenAI Compatible. Do not choose Native Responses.
- Set Base URL to
https://api.guardrelay.ai/v1, API Key togd-…, and Model tomuse-wonder-1.0. - Keep Enable streaming on. Disable Enable R1 model parameters, Use Azure, Enable Reasoning Effort, and custom temperature; leave Custom Headers empty.
- If Include max output tokens is enabled, set Max Output Tokens to
4096as a bounded client example.
| Roo Code field | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://api.guardrelay.ai/v1 |
| API Key | gd-… |
| Model | muse-wonder-1.0 |
| Max Output Tokens | 4096 |
| Final request | POST https://api.guardrelay.ai/v1/chat/completions |
- Open a new task and send:
Read README.md and reply with exactly GUARD_MUSE_OK. Change nothing. - Approve only reading. Then, in the same task, ask for the line without reopening the file.
| Model ID | Input and window |
|---|---|
muse-wonder-1.0 | Text, images, PDF; 1,048,576-token window. |
muse-spark-1.3 | Text, images, PDF; 1,048,576-token window. |
muse-spark-1.2 | Text, images, PDF; 1,048,576-token window. |
- Roo Code itself executes file and terminal tools; it sends the task history to Muse with every request.
- The Guard catalog takes text, images, and PDF as input and returns text. How to send a file or an image depends on the client's interface; this guide does not describe attachments.
- The profile uses OpenAI-compatible Chat Completions. The server prompt is a separate setting of the Muse family: it applies only when the key has the entitlement and the option is on; client setup does not enable it.
| Symptom | What to check |
|---|---|
401 invalid_api_key | The client received the complete active gd-… key; the variable is set in the same process that starts the client. |
404 model_not_found | The model ID is an exact ID from the table above or the short name muse; other spellings are unsupported. |
404 or incorrect path | Base URL is https://api.guardrelay.ai/v1. The client appends /chat/completions; do not add it twice. |
400 invalid_request | The documented API format is selected; extra settings from an older profile were not copied: reasoning, temperature, MCP, and others. |
402 insufficient_credit | Check available balance and unfinished reservations in your Guard account. |
| The model is replaced or the request uses another route | Choose OpenAI Compatible, exact muse-wonder-1.0, and disable Native Responses, R1, and Reasoning Effort. |
This section is for people who already have the Roo Code extension: extension development has ended. For a new installation, use the maintained Cline. In an existing Roo Code installation, configure the Codex family through OpenAI Compatible and Chat Completions. Get your gd-… key from the Guard bot.
For a connection check, use a plain chat without files and ask “Reply: connection established”. To check project work, create a separate guard-test folder in File Explorer or Finder, open it in your editor, and create README.md containing “This is a Guard test project”. The file-reading example below uses that folder.
| Model ID | Name |
|---|---|
gpt-6-astra | GPT-6 Astra |
gpt-6.1-sol | GPT-6.1 Sol |
gpt-6-sol | GPT-6 Sol |
gpt-6-luna | GPT-6 Luna |
gpt-5.6-sol | GPT-5.6 Sol |
gpt-5.6-terra | GPT-5.6 Terra |
gpt-5.6-luna | GPT-5.6 Luna |
- Open VS Code and a test folder without secrets and with a
README.mdthrough File → Open Folder. Open the Roo Code panel from the sidebar icon and select the settings gear. - Choose OpenAI Compatible under API Provider, not Anthropic, OpenAI, or Native Responses.
- Set Base URL to
https://api.guardrelay.ai/codex/v1. - Paste your
gd-…key from the bot into API Key; do not put it in task text or a project file. - Select or enter the exact
gpt-6-astravalue under Model. Turn off Enable Streaming, Include Max Output Tokens, R1 format, Use Azure, Set reasoning level, and Advanced → Use custom temperature; leave Custom Headers empty. See Models for other exact IDs and effort levels. - Save the settings, open a new Roo Code task, and send:
Read README.md and explain in two sentences what this test project does. Do not change anything.If Roo Code asks for permission, allow only a read of that file. - Continue in the same task with:
Name one fact from README.md that supports your explanation.Create a new task for another question.
| Roo Code field | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://api.guardrelay.ai/codex/v1 |
| API Key | gd-… |
| Model | gpt-6-astra |
| Final request | POST https://api.guardrelay.ai/codex/v1/chat/completions |
After checking file reading, you can request a change: “Add a Test section and one descriptive line to README.md.” Review the proposed edit and approve it if it matches your task. Command, file and tool permissions are controlled by the client; connecting Guard does not require granting unrestricted access to your whole computer.