Continue is an open-source AI code assistant and agent framework originally developed by the Continue team (continuedev) and recently acquired by Cursor. It was designed to provide developers with modular, transparent AI capabilities directly within their development environments and workflows, emphasizing a developer-in-the-loop paradigm over opaque full automation. In standard workflows, Continue operates via editor extensions and underlying CLI/agent utilities that allow developers to connect any LLM—including OpenAI, Anthropic, Mistral, or self-hosted models via Ollama and LM Studio. It provides context indexing, inline edits, chat-based refactoring, and customizable slash commands to orchestrate repository-aware generation and debugging tasks using user-supplied API keys or local weights. What historically distinguished Continue from proprietary alternatives is its completely open-source architecture and flexibility. Developers have granular control over prompt templates, custom context providers, and model selection for autocomplete versus chat, preventing vendor lock-in and keeping telemetry and code indexing transparent. Following its acquisition by Cursor, the core Continue repository remains publicly accessible under open-source licensing. Developers can continue to run, fork, and self-host the codebase for individual or enterprise setups without mandatory subscription fees.
Open-source codebase — Fully open-source platform licensed under Apache 2.0, allowing complete code inspection, custom forks, and community modifications. - Bring-your-own-model architecture — Supports connections to remote APIs (such as Anthropic and OpenAI) as well as locally hosted models via Ollama and vLLM. - Custom context providers — Allows integrating codebase indexes, documentation, terminal outputs, and custom data sources into LLM prompts. - Inline edits and side-by-side chat — Supports conversational code generation alongside direct multi-line diff generation and refactoring. - Extensible slash commands — Custom shortcuts to automate recurring developer workflows such as writing tests, generating documentation, or reviewing diffs.
Use cases
Use Case: Offline and Air-Gapped Code Assistance Developers working under strict privacy constraints can pair the open-source Continue harness with local runtimes like Ollama. This enables context-aware chat, autocompletion, and refactoring without sending proprietary source code over public internet endpoints.
Use Case: Custom Context and Repository Indexing Engineering teams can configure Continue to query local git diffs, issue trackers, and internal technical documentation. When asking the assistant to resolve bugs or implement features, the agent automatically pulls relevant files and reference material into the prompt context.
Use Case: Multi-Model Workflow Orchestration Developers can configure lightweight, fast models for latency-sensitive autocompletion tasks while routing complex multi-file refactoring and reasoning queries to larger frontier models, optimizing both cost and latency.
Who it is for
Continue is suited for developers and teams seeking an open-source, vendor-agnostic coding assistant where model endpoints, data privacy, and prompt pipelines must be fully configurable. It is less ideal for teams looking for an out-of-the-box, zero-configuration SaaS agent that requires no setup or API key management.
Install guide
Installation Continue can be installed as an extension within compatible editors (VS Code / JetBrains) or configured via its open-source repository: bash # Clone the open-source repository git clone https://github.com/continuedev/continue.git cd continue # Install dependencies and build locally npm install npm run build
First Run Configure your model provider in the Continue configuration file (located at ~/.continue/config.json): json { "models": [ { "title": "Ollama - Llama 3", "provider": "ollama", "model": "llama3" }, { "title": "Claude 3.5 Sonnet", "provider": "anthropic", "model": "claude-3-5-sonnet-20241022", "apiKey": "YOUR_ANTHROPIC_API_KEY" } ], "tabAutocompleteModel": { "title": "Ollama - StarCoder2", "provider": "ollama", "model": "starcoder2:3b" } }
Example Session Interact with the assistant using context providers and slash commands in your development session: text /edit @src/utils/parser.ts Refactor parseConfig to handle malformed JSON gracefully and return a Result type instead of throwing.
Agentless — Agentless is an open-source automated software development framework developed by OpenAutoCoder. Unlike complex multi-agent architectures that rely on autonomous exploration…
agentdiff — agentdiff is an open-source, Rust-based provenance tracking tool designed to establish Git-native attribution for AI-generated code. Created by Prakhar (codeprakhar25),…
Amp — Amp (developed by AmpCode) is an autonomous coding agent platform and remote development harness designed to execute software engineering tasks…
Kiro — Kiro is an agentic engineering platform and CLI coding agent built and operated by AWS. It extends beyond standard code-completion…
Cline — Cline is an open-source autonomous coding agent runtime maintained by Cline Bot Inc. and an active open-source community. Available as…
OpenHands — OpenHands (formerly OpenDevin) is an open-source autonomous AI software development agent and orchestration platform. It provides an extensible runtime environment…