Twenty 2.0 Launches Open-Source CRM Platform for Developers

Twenty 2.0 ships an open-source CRM built entirely around code-defined data models, workflows, and layouts, with native AI agent support and self-hosting included by default.

The CRM market has handed developers a familiar two-option menu for years, and neither option is good. Buy from a legacy vendor, spend roughly six months wrestling a configuration UI into something resembling your actual business requirements, and end up with a system that works until the moment you need it to do something the vendor didn’t anticipate. Or build something custom, which sounds empowering right up until month three, when your engineering team realizes it’s now a CRM company with a side business. Twenty 2.0 is positioning itself as a third option: an open-source foundation the company explicitly describes as “fully yours,” extended through the same SDK and pull request workflow developers already use for everything else.

That’s the pitch. It’s direct enough that you might underestimate it at first read.

The architecture behind Twenty 2.0 puts everything configurable into code. Custom objects, data models, page layouts, navigation structure, serverless functions, widgets, automated workflows. None of it lives inside an admin panel that only a specific category of power user can navigate. It’s all defined through the SDK and versioned the same way any other application code gets versioned. That distinction matters more in 2026 than it would have in 2022, because the development workflow has changed in the intervening four years. A developer who’s using Cursor or similar AI-assisted coding tools to handle significant chunks of boilerplate isn’t going to be thrilled by a CRM that routes customization through a separate, non-code interface. When the company calls it an “AI-friendly SDK,” that’s not marketing noise. It’s describing a real compatibility with how a lot of engineers are actually working right now.

“Build your Enterprise CRM with an AI-friendly SDK.” That’s the explicit positioning from Twenty, and it’s woven through the full product details posted on Product Hunt.

The self-hosting default deserves more attention than it typically gets in coverage of this kind of product. For companies operating under serious compliance constraints, whether that’s HIPAA, GDPR, financial services regulation, or anything else that makes data residency a legal question rather than an IT preference, the question of where customer data actually lives isn’t a preference. It’s a requirement. Self-hosted by default means that question has an answer on day one. “Fully yours” is doing actual work as a product description, not just brand copy.

The market backdrop for this launch is worth understanding in some detail. Twenty 2.0 arrived when the software development tools market was valued at $8.67 billion in 2026, according to figures from Business Research Insights. That same market is projected to reach $33.9 billion by 2035, at a compound annual growth rate of 14.5%. Those are solid if not spectacular numbers. What’s less moderate is the AI-native subset of developer tools, where the generative AI segment was tracking toward an 84.4% CAGR through 2028, according to reporting from Medium’s tech coverage. Build a CRM with an AI-native SDK in 2022 and you’re building ahead of the curve. Do it in 2026, when AI-assisted coding tools are part of a large share of professional development workflows, and you’re building for the current curve. That timing distinction isn’t trivial.

The product’s standing in the open-source community has real roots. Twenty CRM has logged more than 20,000 GitHub stars, which isn’t a performance metric, but it does indicate sustained developer attention over time. On Reddit, particularly in self-hosted communities, the conversation around Twenty has been consistent and relatively favorable, with one recurring complaint finally addressed in this version: you can now build custom applications on top of the CRM without forking the codebase.

Forking to customize is a trap. It feels like freedom when you do it, and it feels like a maintenance obligation for every subsequent year. Every time the upstream project ships a meaningful update, teams that forked face the same question: do we merge the changes, and if so, what breaks? Twenty 2.0’s SDK approach means custom apps and extensions live outside the core, updated independently, compatible with upstream changes by design. That’s a structural improvement over prior versions, not just a feature addition.

“We’re launching Twenty 2.0 and it’s so big we’ll need four days to cover it,” the company said during the product announcement, which suggests the release contains considerably more than the headline feature list implies. The four-day rollout framing is unusual enough to be worth noting. Most software releases ship a changelog. Rolling out a major version over four days of coverage implies either a genuinely large feature surface or a deliberate effort to sustain launch momentum. Probably both.

The licensing structure is worth understanding for anyone evaluating this for enterprise use. Twenty is released under the GNU Affero General Public License, version 3.0. AGPL is a strong copyleft license with specific implications for network use. If you modify AGPL-licensed software and run it as a service, you’re obligated to make the source of your modifications available. For internal enterprise deployments, this is often a non-issue. For companies building commercial services on top of Twenty, it’s a license question that needs a legal answer before the infrastructure decision is made.

The version jump to 2.0 also signals something about where the project is in its development cycle. The earlier versions clearly served as market validation. Twenty built a developer community, accumulated those GitHub stars, and got enough real-world deployment feedback to know what needed changing. The SDK, the no-fork extensibility, the AI-friendly positioning, these aren’t guesses about what the market wants. They’re responses to what a substantial developer user base asked for.

Enterprise CRM is a crowded category that has been dominated by legacy players for long enough that new entrants mostly come in around the edges rather than challenging the core directly. Twenty 2.0 isn’t quite doing that. An open-source CRM where data stays on your own infrastructure, customization happens through standard development workflow, and AI tools can write the SDK code alongside the rest of your application code, that’s a direct argument against the opinionated, cloud-hosted, UI-configured products that define the legacy end of the market. Whether it converts enterprise buyers at scale is a separate question from whether the argument is coherent. The argument is coherent.

The developer tools market’s trajectory through 2035 gives the underlying bet some structural support. If the market reaches $33.9 billion as projected, and the AI-native developer tooling category continues growing at rates that make 14.5% look conservative, the addressable market for an AI-friendly CRM SDK looks considerably larger in 2028 than it does today. That’s the thesis, and it’s not an unreasonable one.

Twenty 2.0 is available now, licensed under AGPL 3.0, with full product details on Product Hunt.