Workshops Use Cases Architecture Projects Blog
About Contact Offers Talks
Home›AI tools›ContinueDeutsch
continue

Continue: one config file, and you choose every model

Continue is an open-source coding assistant under Apache 2.0 for VS Code, JetBrains and the terminal. Cursor acquired it in 2026, and its repository is now read-only.

Book a free 30-min call →Strands Agents Workshop →

Continue sorts itself around one configuration file. In config.yaml you list your models and give each one roles: chat, edit, apply, autocomplete, embed and rerank. A small model can serve autocomplete while a large one handles chat.

Which model runs and whose account serves it are both your choice. Continue ships no model of its own. It connects to around thirty providers, to local runners such as Ollama, LM Studio and llama.cpp, and to any OpenAI-compatible server you host. This is the open route: it still works, but nobody maintains it any more. Cursor acquired Continue, the final release came out in June 2026, and the GitHub repository is read-only.

From my course slides (MASE-3, FH Burgenland), read 8 September 2026. Some details on the slide have changed since; the text on this page is current as of September 2026.
From my course slides (MASE-3, FH Burgenland), read 8 September 2026. Some details on the slide have changed since; the text on this page is current as of September 2026.[1]

The family

Continue for VS Code[2]

VS Code extension
Extension with Agent, Chat, Edit and Autocomplete. The final 2.0.0 release is on the VS Code Marketplace and Open VSX.

Continue for JetBrains[2]

JetBrains plugin
Plugin for JetBrains IDEs. The README now recommends the CLI instead.

Continue CLI (cn)[3]

CLI
Coding agent in the terminal, published on npm as @continuedev/cli.

config.yaml[4]

local file
The file that holds your models, roles, rules, prompts and MCP servers. Continue creates it on first use in ~/.continue.

Continue Hub[5]

retired
No longer part of the product. The final release switched to explicit model definitions with no Hub references.

Plans

There is no plan to choose. As of September 2026, Continue exists only as open-source code.

PlanForWhat changes
Open source (Apache 2.0)[6]Anyone who installs the extension or the CLI and brings their own models.No paid plan, no admin console and no vendor service are documented. The repository is read-only and gets no more updates.

Which model runs

You choose every model in config.yaml and assign it to roles. Continue never picks for you. As of September 2026, the docs list:

Anthropic, OpenAI, Google Gemini, Mistral, DeepSeek, xAI[7]Hosted model vendors, used with your own API key.
Amazon Bedrock, Azure, Vertex AI, SageMaker[7]Models in your own cloud account and region.
OpenRouter, Groq, Together AI, DeepInfra, HuggingFace and others[7]Hosting services and gateways for many models.
Ollama[7]Local models. Covers all roles, including autocomplete.
LM Studio, llama.cpp, LlamaStack, llamafile[7]Other local runners on your own machine.
HuggingFace TGI, vLLM, SkyPilot, Anyscale, Lambda[8]Self-hosted servers for a team.
QwenCoder2.5 (1.5B, 7B), Codestral, Mercury Coder[9]The docs' suggestions for the autocomplete role.

What you can configure

Model roles[10]Each model lists its roles: chat, edit, apply, autocomplete, embed, rerank. Autocomplete needs a fill-in-the-middle model; "even a 3B parameter model can perform well".
Your own endpoint[8]For an OpenAI-compatible server, use the openai provider and point it at your server. The self-host guide calls the field baseUrl, the config.yaml reference calls it apiBase.
Authentication to your server[8]API key, custom headers, client certificates or a private CA bundle.
Rules, prompts, MCP servers[11]Defined in the same config.yaml as the models.
Offline use[12]Install the .vsix from GitHub, switch off "Allow Anonymous Telemetry" and set a local model as default.
Older config formats[4]config.json, .continuerc.json and config.ts are deprecated. config.yaml replaces them.

What happens to your data

Where prompts go[7]To the provider or server you configure, under that provider's terms. Continue has no model service of its own.
Telemetry[2]The final 2.0.0 release removed anonymous telemetry and sign-in. Older versions sent anonymous telemetry unless you switched it off.
Local and self-hosted[12]With a local runner or your own server, code and prompts stay on hardware you control.

Europe: GDPR, data residency and the AI Act

DPA / GDPR terms[7]Continue publishes no DPA today. The contract that matters is the one with the model provider you connect.
Data residency in the EU[7]It depends on where you point Continue: your own device, your own server, or a cloud region you choose in Bedrock, Azure or Vertex AI.
AI Act[13]Continue ships no model. Neither Continue Dev nor Anysphere (Cursor) is on the EU Code of Practice signatory list; check the provider of the model you use.
Maintenance[2]The repository is read-only, so there are no further security fixes from the vendor. Your organisation carries that risk if it keeps using Continue.

What the AI Act and the GDPR ask of you as a user →

Hybrid: the same models in your own account

This is the open route: the same config file points each job at a model on your own machine, on your own cluster, or in your own cloud account and region. It still works, but it is no longer maintained.

Your own machine[12]Ollama, LM Studio, llama.cpp, LlamaStack or llamafile. Offline use is documented.
Your own cluster[8]Serve open models with HuggingFace TGI, vLLM, SkyPilot, Anyscale or Lambda and point Continue at them.
Your own cloud account[7]Amazon Bedrock, SageMaker, Azure and Vertex AI are providers, so inference runs in your account and in the region you pick.
Mix per role[10]A small local model for autocomplete and a large hosted one for chat, set in the same file.
Limit: no longer maintained[2]The code stays available under Apache 2.0, but "the continuedev/continue repository is no longer actively maintained and is read-only".
The same finished product, in three configurations: consumer, business, or on a cloud account you already have.
The same finished product, in three configurations: consumer, business, or on a cloud account you already have.[14]
Recently changed
  • Cursor acquired Continue. The final 2.0.0 release removed telemetry and sign-in, and the repository became read-only (19 Jun 2026).
  • The final releases dropped Continue Hub references. Configs now need explicit model definitions (15 Jun 2026).

What would help you next?

Learn to decide

The course “From Managed to Hybrid” shows where AI should run: what each layer costs, what you operate and where your data has to stay. In preparation, with a free live session first.

Join the course waitlist →
Linda MohamedLinda Mohamed
AI and cloud consultant, AWS Community Hero, Vienna

Roll it out with me

Tell me which tools your teams use or want. In 30 minutes we look at plans, settings, data rules and whether AWS funding can cover part of a pilot. How AWS funding works →

Book a free 30-min call →15-min question slot

You already know what you need

Book a workshop directly:

ai-solutions.wiki

Read more on ai-solutions.wiki

My open knowledge base, with longer comparisons and sources.

The technology underneath: AI technologies explained →

Sources

  1. Linda Mohamed, course slides “The projects”, appendix layer 0, MASE-3, FH Burgenland, September 2026
  2. Continue for VS Code, github.com. https://github.com/continuedev/continue
  3. Continue CLI (cn), docs.continue.dev. https://docs.continue.dev/
  4. config.yaml, docs.continue.dev. https://docs.continue.dev/customize/deep-dives/configuration
  5. Continue Hub, github.com. https://github.com/continuedev/continue/releases/tag/v1.2.23-vscode
  6. Open source (Apache 2.0), continue.dev. https://continue.dev/
  7. Anthropic, OpenAI, Google Gemini, Mistral, DeepSeek, xAI, docs.continue.dev. https://docs.continue.dev/customize/model-providers/overview
  8. HuggingFace TGI, vLLM, SkyPilot, Anyscale, Lambda, docs.continue.dev. https://docs.continue.dev/guides/how-to-self-host-a-model
  9. QwenCoder2.5 (1.5B, 7B), Codestral, Mercury Coder, docs.continue.dev. https://docs.continue.dev/customize/model-roles/autocomplete
  10. Model roles, docs.continue.dev. https://docs.continue.dev/customize/model-roles/00-intro
  11. Rules, prompts, MCP servers, docs.continue.dev. https://docs.continue.dev/reference
  12. Offline use, docs.continue.dev. https://docs.continue.dev/guides/running-continue-without-internet
  13. AI Act, digital-strategy.ec.europa.eu. https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai
  14. Linda Mohamed, course slides “Start here”, MASE-3, FH Burgenland, September 2026
  15. vLLM, ai-solutions.wiki. https://ai-solutions.wiki/tools/vllm/
  16. Managed vs reserved vs self-hosted inference, ai-solutions.wiki. https://ai-solutions.wiki/comparisons/managed-vs-reserved-vs-self-hosted-inference/
  17. Release v2.0.0-vscode (19 Jun 2026). https://github.com/continuedev/continue/releases/tag/v2.0.0-vscode
  18. Apache 2.0 licence. https://raw.githubusercontent.com/continuedev/continue/main/LICENSE
  19. npm @continuedev/cli. https://www.npmjs.com/package/@continuedev/cli

Checked against the vendors' own pages in September 2026. Plans, models and features change often: check the linked pages before you decide.