Continue - Build & Share AI Coding Assistants

Continue Launches Version 1.0 with $3 Million Seed Funding

A recently established company is focused on empowering developers to build personalized, context-aware coding assistants.

These assistants are designed for compatibility with various models and smooth integration into existing development workflows.

Company Origins and Growth

Continue was founded in June 2023 by Ty Dunn, serving as CEO, and Nate Sesti, the CTO.

The company, a graduate of the Y Combinator program, has rapidly gained traction within the developer community.

Over the last two years, Continue has accumulated over 23,000 stars on GitHub and fostered a community of 11,000 members on Discord.

New Product Release and Funding

Building upon its initial success, Continue is now releasing version 1.0 of its core product.

This launch is accompanied by a new round of seed funding totaling $3 million.

The funding will be utilized to further develop the platform and expand its capabilities for developers seeking to create tailored coding assistants.

The Rise of AI Coding Assistants

The introduction of Continue arrives during a significant surge in the availability of AI-powered coding assistants. Prominent examples include GitHub Copilot and Google’s Gemini Code Assist. Furthermore, newer companies like Codeium and Cursor have secured substantial investment.

Continue positions itself as a premier open-source AI code assistant. It boasts the capability to integrate with any AI model. Teams are also empowered to incorporate their own specific data through connections to platforms such as Jira and Confluence.

By linking models and relevant context, developers can construct tailored autocomplete and chat functionalities directly within their coding environments. Autocomplete delivers real-time code suggestions during typing. The chat feature allows developers to pose questions regarding specific code segments.

Code modification is also streamlined. Users can describe desired changes, and the system will implement them.

Today’s announcement includes the initial “major” release of Continue’s open-source extensions. These are available for both VS Code and JetBrains IDEs.

“This demonstrates to businesses that the project is stable and reliable for building upon,” explained Dunn in a conversation with TechCrunch.

In addition, Continue is unveiling a new hub. This hub functions similarly to Docker Hub, GitHub, or Hugging Face. It serves as a central location for developers to create and distribute custom AI code assistants.

The hub also features a registry for defining and managing the components used in these assistants. At its launch, the hub provides pre-built AI coding assistants.

It also includes “blocks” from verified partners like Mistral (with its Codestral model), Anthropic (Claude 3.5 Sonnet), and DeepSeek-R1 via Ollama. However, contributions of blocks and assistants are open to all developers and vendors.

A “block” can encompass various elements. These include AI models and their locations, customization rules for the assistant, context providers like Jira or Confluence, pre-written prompts for complex instructions, documentation sources (such as Angular or React), data transmission for analytics, and MCP servers for building and sharing language model tools.

A Collaborative Approach to AI Development

The core concept driving this new platform is the expectation that most users will not necessitate extensive personalization. Instead, they will primarily focus on implementing small adjustments to existing coding assistants or pre-built blocks within the hub.

This naturally leads to the consideration of incentives for developers to create and distribute their customizations. The motivation mirrors that found in established open source communities. Several of the initial partners are the very organizations responsible for the foundational tools and AI models – including companies like Mistral and Anthropic – positioning Continue’s hub as a strategic location for fostering developer relationships.

Embracing the Open Source Ethos

At its foundation, Continue champions an “open source ethos.” If customizations are developed for internal use within a company, the logical extension is to share them with the broader developer community. Continue aims to differentiate itself from providers offering proprietary, “black box” AI assistant solutions.

“Our vision is a central hub for the entire ecosystem to collaborate,” explained Dunn. “Rather than each entity constructing its own isolated, closed-source AI code assistant, we propose an open architecture enabling collective development of the essential components needed for personalized experiences.”

Dunn describes this as cultivating a “culture of contribution,” encouraging developers to innovate and build customizations that benefit the entire user base.

“Continue 1.0 empowers this culture of contribution, allowing developers to create and share custom AI code assistants,” Dunn stated. “This registry will serve as a discovery resource both within and across organizations, evolving alongside the development of blocks and enhanced AI developer tools.”

Data Control and Transparency

A key consideration is data control. Generic, “one-size-fits-all” platforms often leverage aggregated developer usage data to refine their services. This practice has sparked debate, as exemplified by criticisms leveled against GitHub Copilot, accused of capitalizing on the contributions of open source software developers.

Continue prioritizes giving companies greater autonomy over their data, allowing them to determine the extent of data sharing.

“When utilizing Continue, your data remains under your control,” Dunn emphasized. “Organizations can consolidate their data for all developers in a single, secure location. This is not feasible with the ‘one-size-fits-all’ approach, where the business model relies on data extraction for platform improvement.”

• Coding assistants are easily customizable.
• The platform fosters a culture of contribution.
• Organizations maintain complete data control.

A Promising Business Model

Continue is a relatively new company, yet it reports collaboration with several prominent businesses during its development phase. These include Ionos, an existing Continue client, alongside Siemens and Morningstar.

While focusing significantly on larger enterprises, Continue aims to serve developers of all sizes – from independent freelancers and small teams to expansive organizations. This strategy defines its revenue model; the newly launched hub offers a complimentary solo tier.

Organizations requiring enhanced data control can subscribe to access advanced administration, governance, and security features. According to Dunn, the solo tier is adequate for individual developers seeking customization.

As freelancers or small teams expand and require governance, they can transition into paying customers. The free solo tier provides three distinct “visibility” levels for contributions.

Developers can choose to keep their work private, share it within a team, or make it publicly available. Technically, the solo tier supports team collaboration, though it lacks features typically needed by teams.

The “teams” tier introduces enhanced “multi-player” functionality, including administrative controls to manage access to blocks and assistants. The enterprise tier further elevates data security and governance options.

It provides more precise control over utilized blocks, models, versions, and vendors. Administrators can also oversee credential security, data destinations, and access a comprehensive audit log detailing developer activity.

Continue had already secured $2.1 million in funding after completing the Y Combinator program in late 2023. Now, the company has raised an additional $3 million through SAFEs, led by Heavybit, a venture capital firm specializing in developer tools.

The majority of this new funding will be allocated to software engineering salaries, with plans to at least double the current team of five employees. Strategic cost management is a key element of their approach.

“Leveraging open source for distribution allows us to maintain low costs, reducing the capital expenditure needed compared to competitors,” Dunn explained. This approach allows for efficient growth and development.