NocoBase is not just adding AI labels to a no-code product. It is shipping through multiple release tracks while rebuilding pieces of the product around AI employees, agent access, and a v2 client. ToolVitals sees 30 release events in 30 days, 30 GitHub releases in 90 days, a 100 shipping score, and a 97 ToolVitals score.
That is a lot of motion for a business-app builder.
The official site frames NocoBase as an open-source AI + no-code platform for building business systems. The pitch is specific: AI works on top of existing infrastructure and a WYSIWYG no-code interface, rather than generating a fresh app from scratch. The site also calls out Claude Code, Cursor, Codex, OpenCode, and TRAE as agent paths into the product.
The release data backs up that positioning. Recent releases include AI employee fixes, an LLM Base URL improvement, workflow rerun APIs, client-v2 security settings, plugin manager work, Office File Previewer migration, notification plugin migration, and FlowEngine-related fixes. This is not one isolated AI feature. It looks like a broad v2 migration with AI and workflow features moving in parallel.
The signal: three trains at once
NocoBase publishes stable, beta, and alpha updates close together. On May 29, ToolVitals captured v2.0.60, v2.1.0-alpha.45, v2.1.0-beta.38, and v2.1.0-beta.40. The official release notes page also describes three branches: main for stable installs, next for beta testing, and develop for alpha iteration.
That release structure matters. No-code business systems are sticky, permission-heavy, and workflow-heavy. Teams do not want surprises in production. NocoBase seems to be separating production stability from fast v2 testing, which is the right shape for this category.
The interesting bet is not simply “AI builds apps.” The bet is that agents and humans can share the same business system surface. The website says agents get CLI and skills, while people get visual configuration. Recent releases add or migrate security menus, password policy, plugin management, workflow timeout controls, public forms, and AI employee behavior.
That is boring plumbing, in a good way. AI features for enterprise apps are useless if permissions, workflows, migrations, and admin screens are weak.
What ToolVitals cannot tell from this data
ToolVitals can see public release cadence, GitHub stars, score signals, SSL and uptime checks, and first-party positioning. It cannot measure code quality, production reliability inside customer deployments, user satisfaction, revenue, support load, or whether AI employees work well on real business tasks.
The openness field in the ToolVitals payload marks NocoBase as OSI-approved OSS with an Apache-2.0 license signal. During this check, GitHub API metadata returned NOASSERTION for the repository license, so license prose should stay tied to the ToolVitals openness data unless a maintainer points to a clearer current license page.
The release notes also mention commercial license components and license settings work. That does not change the ToolVitals openness classification, but it does mean buyers should check feature boundaries before assuming every plugin or deployment path is covered the same way.
How it compares
NocoBase is smaller than LangChain by stars, 22,609 versus 138,238 in the ToolVitals payload. But its 30 release events in 30 days beat LangChain’s 24 in the same window.
It matches Composio on 30 release events in 30 days and a 100 shipping score, while sitting at 22,609 stars versus Composio’s 28,572. Compared with PostHog, NocoBase ships fewer counted release events, 30 versus 42, but PostHog is marked open core while NocoBase is classified by ToolVitals as OSI-approved OSS.
Those comparisons do not say NocoBase is better. They say it is operating at the release tempo of much louder developer tools.
Recommendation
If your team builds internal business apps and wants AI assistance without giving up visual configuration, evaluate NocoBase now. Focus the trial on permissions, workflows, plugin boundaries, and the v2 client path. The release cadence is strong enough to justify attention, but the real test is whether its AI employees and no-code controls survive your messy business rules.