Integrate RodiumAi manually with Continue.dev in VS Code

Why Continue.dev + RodiumAi?

Continue.dev is a free, open-source AI coding assistant for VS Code. It adds a chat sidebar, inline Edit, and Agent modes directly in your editor — like having GPT, Claude, or Gemini inside your IDE.

RodiumAi gives you access to 50+ frontier models (GPT-5.4, Claude Sonnet 4.6, Gemini 3.5 Flash, DeepSeek, Kimi, Grok, and more) through one API key and one wallet (RODI credits). No need to juggle separate OpenAI, Anthropic, or Google accounts.

This guide walks you through the full setup in about 5 minutes:

1. Create your RodiumAi account and API key

2. Create a .env file in the .continue folder

3. Copy the model list from GitHub into config.yaml

4. Pick a model and start coding

Branding: write RodiumAi (not RodiumAI).

Before you start

You will need:

• VS Code with the Continue extension installed

• A RodiumAi account — if you don't have one yet, sign up free at rodiumai.io

• An API key — create one in Dashboard → API Keys (starts with rd_sk_prod_…; shown once at creation — copy it immediately)

• A positive RODI balance — if your wallet is empty, recharge with Mobile Money from the dashboard before your first request

Tip: create one API key per project so usage tracking stays clean in Dashboard → Usage.

Step 1 — Create your API key on RodiumAi

If you already have a key, skip to Step 2.

1. Go to rodiumai.io and sign in (or create an account).

2. Open Dashboard → API Keys.

3. Click Create API key, give it a name (e.g. continue-vscode), and copy the full secret (rd_sk_prod_…).

4. Store it somewhere safe — RodiumAi shows the full key only once. If you lose it, regenerate a new key from the same page.

Not sure which model to use later? Browse the live catalogue at rodiumai.io/models — every slug there works with Continue.

Step 2 — Open the .continue folder and create .env

Continue stores its configuration in a hidden folder on your machine. Your API key goes in a .env file there — Continue does not read Windows environment variables or PowerShell $env:….

Windows

1. Press Win + R to open the Run dialog.

2. Type %USERPROFILE%\.continue and press Enter.

   • File Explorer opens the Continue folder (usually C:\Users\<YourName>\.continue).

3. If the folder does not exist, create it:

   • Go up one level (your user folder), right-click → New → Folder, name it .continue, then open it.

4. Create the .env file:

   • Right-click inside the folder → New → Text Document.

   • Name it exactly .env (remove .txt if Windows adds it — confirm when Windows warns about changing the extension).

   • Or open Notepad, paste the line below, then File → Save As, set filename to .env, choose All files (.), and save inside .continue.

5. Open .env with Notepad and add one line (no quotes, no spaces around =):

RODIUM_API_KEY=rd_sk_prod_paste_your_full_key_here

Replace rd_sk_prod_paste_your_full_key_here with your real key from Step 1. Save and close.

Never commit this file to git. It stays only on your machine.

macOS / Linux

Open a terminal and run:

mkdir -p ~/.continue
nano ~/.continue/.env

Paste the same line with your key, save (Ctrl+O, Enter), and exit (Ctrl+X).

Step 3 — Copy the model list from GitHub

RodiumAi maintains a ready-made config.yaml with every coding-capable model from the live API — no need to type slugs one by one.

ResourceLinkGitHub repogithub.com/Rodium-AI/continue.dev.configRaw config (copy this)rodiumai-models.config.yaml

1. Open the raw config link above in your browser.

2. Press Ctrl + A (select all), then Ctrl + C (copy).

   • The file uses a YAML anchor (rodium_defaults) so apiBase, apiKey, and useResponsesApi are defined once — you never repeat your connection settings on every model.

The config already points to https://api.rodiumai.io/v1 and references ${{ secrets.RODIUM_API_KEY }} from your .env file. You do not paste your API key into config.yaml.

Alternative — download in one command (Windows PowerShell):

Invoke-WebRequest `
  -Uri "https://raw.githubusercontent.com/Rodium-AI/continue.dev.config/main/rodiumai-models.config.yaml" `
  -OutFile "$env:USERPROFILE\.continue\config.yaml"

If you use this command, you can skip the manual paste in Step 4 — but still complete Step 2 (.env) first.

Step 4 — Paste into config.yaml in VS Code

1. Open VS Code.

2. Press Ctrl + Shift + P (Command Palette).

3. Type Continue: Open config.yaml and press Enter.

   • VS Code opens %USERPROFILE%\.continue\config.yaml (or ~/.continue/config.yaml on macOS/Linux).

4. Select all existing content (Ctrl + A) and paste (Ctrl + V) the GitHub config you copied in Step 3.

5. Press Ctrl + S to save.

If the file was empty before, you may have seen only:

name: Local Config
version: 1.0.0
schema: v1
models: []

After pasting, you should see dozens of entries like Claude Fable 5 (RodiumAi), GPT-5.4 (RodiumAi), Gemini 3.5 Flash (RodiumAi), etc., all inheriting shared settings via <<: *rodium_defaults.

Step 5 — Reload Continue

After saving config.yaml:

1. Press Ctrl + Shift + P again.

2. Run Continue: Reload config.

   • Or close and reopen VS Code.

Continue reads the new model list immediately.

Step 6 — Choose a model and start coding

1. Click the Continue icon in the VS Code activity bar (left sidebar) to open the Continue panel.

2. At the top of the panel, open the model dropdown.

3. Pick any RodiumAi entry — for example:

   • Claude Sonnet 4.6 (RodiumAi) — strong balance of speed and quality

   • GPT-5.4 Mini (RodiumAi) — fast and cost-efficient

   • Claude Fable 5 (RodiumAi) — long-horizon agentic coding

4. Open a project folder in VS Code, then:

   • Chat — ask questions about your codebase

   • Edit — select code and describe the change you want

   • Agent — let Continue work across multiple files

Try a first prompt: "Explain what this project does in three bullet points."

If you get a reply, you're live. Check consumption in Dashboard → Usage on RodiumAi.

How the config works (quick reference)

FieldRodiumAi valueprovideropenai (OpenAI-compatible client — RodiumAi is not OpenAI, but Continue speaks this protocol)apiBasehttps://api.rodiumai.io/v1apiKey${{ secrets.RODIUM_API_KEY }} — resolved from your .env filemodelExact catalogue slug, e.g. openai/gpt-5.4useResponsesApifalse — avoids OpenAI /responses errors on GPT-5 / o-series models<<: *rodium_defaultsYAML merge — each model inherits shared connection settings

Want a smaller list? Keep only the models you use in the models: block, or browse rodiumai.io/models and copy slugs exactly (including the vendor/ prefix).

Troubleshooting

SymptomLikely causeFix401 UnauthorizedWrong or missing API keyCheck .env in %USERPROFILE%\.continue\; regenerate your key402 Payment RequiredLow RODI balanceRecharge in Dashboard → BillingModel not in dropdownConfig not reloadedRun Continue: Reload config or restart VS CodeUnknown model / 404Outdated configRe-copy the latest file from GitHubSecret not resolvedTypo in .envVariable must be RODIUM_API_KEY (matches ${{ secrets.RODIUM_API_KEY }} in YAML)Organization must be verifiedOpenAI Responses APIAlready set to useResponsesApi: false in the GitHub config

What's next?

• Recharge RODI anytime via Mobile Money on rodiumai.io/dashboard/billing.

• Compare models and pricing on rodiumai.io/models.

• Track spend per key in Dashboard → Usage.

• Try other tools with the same RodiumAi key — OpenCode, Cursor, Cline: rodiumai.io/insights.

One account. One wallet. Every frontier model — directly in VS Code.