Claix launches an AI document-processing and context API

Summary

Claix is an AI document-processing and context platform for developers, backends, automations, and agents. It converts Excel and CSV files to JSON, as well as text, HTML, XML, PDFs, Word documents, images, and audio through REST endpoints. Developers can query one document by document_id or group documents under a space_id for cross-document comparison, aggregation, and reconciliation. The service can be integrated through REST, the Agent-to-Agent protocol, or MCP, including connections to Cursor, Claude Desktop, and Windsurf. Claix offers permanent, temporary, or immediate-deletion storage modes; it says data is hosted in the EU, encrypted in transit and at rest, isolated by tenant, and not used by Claix or its providers to train AI models. Its schema-based extraction produces typed JSON, while Source Tracing attaches evidence to values such as pages, cells, or text fragments. When a field cannot be grounded, the output can mark it requires_human_revision instead of supplying an unsupported citation. The platform supports multimodal processing of scanned PDFs, spreadsheets, Word files, and images, and claims to reduce prompt-token consumption by more than 80 percent in its described workflow. Pricing is usage-based: the first 100 successful calls are free, then extraction calls cost €0.15 and context calls cost €0.03, with enterprise and self-hosted options also listed.