Databutton Integrates Claude Sonnet 4.5 Into Riff Platform for Internal Tool Development.

Databutton has integrated Claude Sonnet 4.5 into its Riff platform, giving teams access to what Anthropic positions as its most capable coding model for building internal AI tools and workflows.

The integration brings Sonnet 4.5's benchmark performance to Riff's existing no-code builder. Anthropic's model scored 77.2% on SWE-bench Verified, a test that measures how well AI systems can resolve real-world software engineering issues from GitHub repositories. For context, GPT-4o scored 38.8% on the same benchmark when it launched in May 2024.

Databutton positions the update as a way for teams to move past pre-built SaaS tools and monthly subscriptions. The company suggests users can now build AI-powered marketing dashboards, automated call summary systems that post to Slack, or lead enrichment workflows that surface qualified prospects without relying on third-party platforms.

The platform supports long-running tasks, context editing, and memory, making it suited for complex apps and agents that need to maintain state across interactions. This follows Databutton's recent push into vertical-specific automation, where the company has targeted finance and sales teams with workflow tools.

Riff competes in a crowded space alongside tools like Zapier, Make, and n8n, but differentiates by positioning itself as a full development environment rather than just a workflow connector. Where Zapier focuses on linking existing apps, Riff lets users build new interfaces and logic on top of those connections.

Databutton is offering 20 free credits to new users who want to test Sonnet 4.5 within the platform. The company did not disclose per-token pricing or how credits map to API usage, but positioned the offer as a way for teams to prototype internal tools without upfront cost.

The integration adds to Riff's existing model support. Databutton added Google's Gemini 3 models earlier this year, giving users options across providers depending on use case and budget.

Claude Sonnet 4.5 launched in October 2024 as Anthropic's latest iteration in the Sonnet series, designed specifically for coding tasks. The model introduced improvements in code generation, debugging, and multi-step reasoning over previous versions. Anthropic has emphasized agentic capabilities, meaning the model can handle tasks that require planning, iteration, and error correction without constant human supervision.

For Databutton users, this means Riff can now tackle more complex builds autonomously. A team could theoretically describe a procurement dashboard that pulls data from multiple APIs, formats it into tables, and sends alerts when thresholds are hit. The model would generate the necessary code, handle edge cases, and iterate on feedback within the platform's interface.

The update arrives as AI-powered development tools continue to blur the line between no-code and traditional programming. Platforms like Cursor, Replit, and Lovable have all integrated frontier models to let users build software through natural language. Databutton's angle is targeting internal ops teams who need custom tools but lack dedicated engineering resources.

Whether Sonnet 4.5 delivers enough improvement over existing options to change buying behavior remains to be seen. The benchmark scores suggest strong technical performance, but the real test will be whether teams can ship and maintain tools faster than they could with existing platforms or hiring developers outright.