How to use Darkbloom with Continue

At a glance

How do I use Darkbloom with Continue in VS Code or JetBrains?

Add a Darkbloom model to Continue’s config.yaml and keep your key in a .env file. Continue’s docs say the VS Code and JetBrains extensions use the same config file. We ran the VS Code extension, so the steps below describe that one.

1. Get a Darkbloom API key. Open the Darkbloom console and sign in, add credits (the minimum is $0.50, and Darkbloom says credits are non-refundable and never expire), then choose API console in the console’s left menu and Create API key (Darkbloom’s docs call this page Settings → API Keys). The console shows the key only once, so copy it right away. Every key spends from your prepaid balance, so add only what you are willing to spend while you test, and watch the first sessions, because Agent mode sends one request for each step it takes.
2. Install Continue. In VS Code, install Continue from the Marketplace. In a JetBrains IDE, install Continue from its plugin page. The first time Continue runs, it creates the folder ~/.continue (on Windows, %USERPROFILE%\.continue) with a starter config.yaml.
3. Save your key in a .env file. Continue’s extensions cannot read environment variables from your shell (Continue docs), so the key goes in a file. Create ~/.continue/.env (on Windows, %USERPROFILE%\.continue\.env) and put one line in it: the variable name, an equals sign and your key, with no quotes and no spaces.

   ~/.continue/.envCopy

   DARKBLOOM_API_KEY=YOUR_DARKBLOOM_API_KEY

   macOS and Linux, optional: only you can read the fileCopy

   chmod 600 ~/.continue/.env

4. See which models you can use. On macOS and Linux, load the key from that file and ask Darkbloom for its model list, which gives each ID with its context window and prices. On Windows, pick an ID from the table further down this page instead.

   Copy

   set -a; . ~/.continue/.env; set +a
   curl -s https://api.darkbloom.dev/v1/models \
     -H "Authorization: Bearer $DARKBLOOM_API_KEY"

5. Add Darkbloom to config.yaml. Open it at ~/.continue/config.yaml, or from Continue’s panel: click the gear icon, then Configs, then the gear next to your config. The panel has to be wide enough to show its settings, so drag its edge if it says the screen is too small. A new install starts with an empty model list, models: []. Replace the file’s contents with this, or, if you already have models or other settings in it, add only the model entry under models:.

   config.yamlCopy

   name: Darkbloom
   version: 0.0.1
   schema: v1
   models:
     - name: Darkbloom gemma-4-26b
       provider: openai
       model: gemma-4-26b
       apiBase: https://api.darkbloom.dev/v1
       apiKey: ${{ secrets.DARKBLOOM_API_KEY }}
       defaultCompletionOptions:
         contextLength: 131072

6. Choose the model and send a message. Save the file. Continue reloads it by itself within a few seconds, and the model appears in the model menu under the chat box. Pick it and send a message such as Say hello in five words. If you created or changed the .env file after Continue started, open the config menu above the chat box (it shows your config’s name, Main Config on a new install) and choose Reload, because Continue reads .env files when it loads the config.

What does each line of the Continue config mean?

This is the model entry from the steps above, line by line.

Three optional settings are not in that entry. maxTokens, under defaultCompletionOptions, sets the longest answer Continue asks for: the extension asks for 4,096 tokens when you set nothing (seen in its requests), and the CLI sends no limit, so Darkbloom adds 8192 (Darkbloom docs). roles says what the model is used for. Continue’s default is chat, edit, apply and summarize (config.yaml reference), so Darkbloom is not used for tab autocomplete unless you add autocomplete; read the cost section first. capabilities tells Continue what the model can do. It was not needed for gemma-4-26b: Agent mode and tool calls worked without it. If Agent mode is greyed out for another model, add tool_use under capabilities, but only for a model whose Features column below lists tools (Continue docs).

Can I use Darkbloom with the Continue CLI?

Yes. The CLI, cn, reads the same ~/.continue/config.yaml and .env files as the extensions (Continue docs), so after the steps above it already uses Darkbloom.

1. Install the CLI. It needs Node.js 20 or newer (Continue docs). Continue’s CLI quickstart also lists a shell installer for macOS and Linux and one for Windows.

   Copy

   npm i -g @continuedev/cli

2. Run one prompt. In your project folder, -p sends a single prompt, prints the answer and exits:

   Copy

   cn -p "Say hello in five words."

To use another config file, add --config ./my-config.yaml. To let the agent change files without asking, add --auto, which allows every tool, so use it only in a folder you can restore. We ran headless mode (-p) with a local config, and it needed no Continue login. We did not open the interactive screen (cn alone), which Continue’s docs say asks you to log in to Continue or enter an Anthropic key on the first launch.

Which Darkbloom model should I use with Continue?

Use any ID that GET /v1/models returns. On October 23, 2026 it listed 12 models, shown here from the cheapest output price up, in US dollars per million tokens. Check the live prices before you rely on a number; that public list also names models that are not online. A key can be limited to some models, and Darkbloom refuses any other with a 403 (see the errors table).

Add one entry per model under models:, each with its own name and model, and Continue lists them all in its model menu. For Agent mode, choose a model whose Features column lists tools, and set contextLength to that model’s context window.

How much does Continue cost with Darkbloom, and how do I keep it low?

Continue adds no charge. Darkbloom bills the tokens you use from your prepaid balance: (input tokens × input price + output tokens × output price) ÷ 1,000,000, with a minimum of $0.0001 per request. Continue sends more than your message, so a short question is a bigger request than it looks. In Agent mode, the default for new chats, one short message sent 1,985 tokens in and got 8 back, because Continue adds its own instructions and 12 tool definitions. At the October 23, 2026 prices that is worth $0.000085 on gemma-4-26b, so Darkbloom billed the $0.0001 minimum. The same message in Chat mode sent 397 tokens in. A large request, such as 20,000 tokens in and 2,000 out, costs about $0.00128. In Darkbloom’s usage log each of the short messages above was billed $0.0001. One request of our file edit, 2,091 tokens in and 78 out, was billed $0.000104, just over the minimum.

To limit the length of each answer, set maxTokens as described above. Check your real spend in the Darkbloom console, or call GET /v1/payments/usage with your key (Darkbloom docs: usage history).

Where does Continue look for my API key?

Continue swaps ${{ secrets.DARKBLOOM_API_KEY }} for the value of DARKBLOOM_API_KEY that it finds in the places below. Keep the key in ~/.continue/.env, as in the steps above: it worked in the extension and in the CLI.

Continue’s docs list the project’s root .env for the extensions too, but in our test the extension did not read it. Use ~/.continue/.env or <project>/.continue/.env. Keep the key in one place: when two places had a value, the extension and the CLI picked different ones. Continue reads .env files when it loads the config, so after you change one, choose Reload in the config menu. Continue’s docs say to restart the IDE, and Reload was enough in our test.

If you keep the key in a project folder, add the file to .gitignore so it is never committed. Do not paste the key into config.yaml itself if you commit or share that file.

echo ".continue/.env" >> .gitignore

What do the common errors mean?

The extension shows an error in a window titled “Error handling model response”. Its last part, under “View error output”, is Darkbloom’s answer, and the CLI prints the same text as {"status":"error","message":"..."}. These are the messages you are most likely to see. Each was reproduced with Continue 2.1.0 or the CLI, except the empty-balance row.

Is Darkbloom private, and is it ready for production?

Not yet for sensitive work. Darkbloom’s quickstart calls the service “an experimental research prototype” and says not to use it in production applications, and its console labels it Public alpha. Darkbloom says prompts are encrypted to a verified provider Mac. Its security overview says the coordinator re-encrypts each request for the chosen provider and runs in a confidential VM. An independent review by Wavect (September 2026) reads Darkbloom’s technical documentation as saying that both the coordinator and the provider handle your prompt as plaintext inside protected boundaries.

Continue sends your messages, the files and code you attach and, in Agent mode, whatever its tools read from your project, such as the contents of a file, to Darkbloom. Do not send secrets, customer data or code you are not allowed to share.

How this guide was checked

• A paid request from Continue to api.darkbloom.dev worked on October 22, 2026. We ran the steps on this page again as written. Continue extension 2.1.0 in VS Code 1.143 answered a message in Chat mode and in Agent mode and made a file edit through tool calls, and Continue CLI 1.5.47 answered cn -p "Say hello in five words.", all with gemma-4-26b. Darkbloom’s usage log shows those requests billed between $0.0001 and $0.000104.
• The setup steps were compared with Continue’s own docs (linked under Sources) on October 22, 2026. Where the docs and the program differ, this page follows what we saw: the config menu shows your config’s name (Main Config), not “Local Config”, Reload picked up a changed .env file where the docs say to restart the IDE, and the extension did not read a .env file in the project’s root folder where the docs say it does.
• We ran the Continue extension 2.1.0 (the Linux build from Open VSX) in VS Code 1.143, run as a web server and opened in a browser, with a fresh Continue folder. It created the starter config.yaml, loaded the config from this page, reloaded it when the file changed, and answered from Darkbloom in Chat mode and in Agent mode. To see what it sends, we also ran it through a local forwarder that passed each request to Darkbloom unchanged and logged its size and settings. In Agent mode it read a file, edited it through tool calls and showed the change as a diff. We also checked the config menu, the settings page, the model and mode menus and the error windows shown above.
• We ran the Continue CLI 1.5.47 on Linux, installed with npm install: headless prompts (cn -p) with --config and with the default ~/.continue/config.yaml, a chat and an --auto file edit through tool calls. We also checked which .env files and which environment variables it reads.
• The error rows were reproduced with the extension or the CLI against Darkbloom, except the “invalid API key” rows, which used a test server that returns Darkbloom’s real reply for a wrong key. Darkbloom’s own text for an empty balance was not captured, because that needs an empty balance.
• Two details come from Continue’s source code on GitHub, not from a test: the 32,768-token default context length, and the order in which the extension looks for .env files (our tests agreed with that order).
• The model list, context windows, features and prices come from GET /v1/models on October 22, 2026 (it needs a key). The key steps in the console come from the console’s own page text. We tested with a key that already existed, so we did not click through key creation ourselves.
• Our test machine adds that key to requests to Darkbloom on their way out, so in the runs against Darkbloom the .env file held this page’s placeholder line. That Continue sends the value from a .env file was checked separately, with a local test server that prints the Authorization header it receives.
• Not run: the JetBrains extension, the VS Code desktop app, the Windows commands and paths, tab autocomplete, Edit and Apply, image attachments, codebase indexing and the CLI’s interactive screen.
• Our Darkbloom link was opened in a clean browser profile on October 22, 2026. It saved the referral code and the page said the code “is saved and will be applied after you sign in”. We did not create an account to test the sign-in step itself.

About this guide

BloomGauge is a free Mac app for people who run Darkbloom providers, so we benefit when Darkbloom has more paying demand. We are independent of Darkbloom and not endorsed by it. Our Darkbloom sign-in links carry our referral code, BLOOMGAUGE. If you sign in to Darkbloom after opening one, Darkbloom attaches the code to your account and pays us a share of your eligible token spend. Darkbloom funds that reward, so it does not change your price.

Sources

• Continue docs: OpenAI and OpenAI-compatible models
• Continue docs: config.yaml reference
• Continue docs: FAQs, including secrets and .env files
• Continue docs: how to understand configuration
• Continue docs: model capabilities
• Continue docs: CLI quickstart
• Continue docs: CLI configuration
• Continue on the VS Code Marketplace
• Continue source code: how secrets are found (LocalPlatformClient.ts)
• Continue source code: default context length and token limit (constants.ts)
• Darkbloom docs: quickstart
• Darkbloom docs: chat completions
• Darkbloom docs: authentication and API keys
• Darkbloom docs: add credits
• Darkbloom docs: usage history
• Darkbloom live prices (GET /v1/pricing)

Found something out of date? Tell us in #bloomgauge on the Darkbloom Slack or at Support.