Cradle

Cradle is an AI-powered protein engineering platform designed specifically for scientists in pharma, industrial biotech, foodtech, and adjacent fields. By combining advanced machine learning with an intuitive interface, Cradle enables R&D teams to produce successful protein candidates 2 to 12x faster than traditional workflows, shortening timelines across discovery and optimization programs while improving overall pipeline success rates.

Used by leading R&D teams and continuously validated in Cradle's own Amsterdam wet lab, the platform takes an agnostic approach to protein modalities and data types — covering enzymes, peptides, and single- or multi-chain antibodies — so that if a property can be measured, Cradle can help optimize it. Scientists consistently rate Cradle's AI-engineered candidates 8 out of 10, reflecting sequences that perform in real-world wet lab conditions, not just in silico.

Core AI Capabilities

• Simultaneous multi-property optimization: Unlike traditional sequential approaches, Cradle's AI engine understands how target properties interact and optimizes them in parallel — balancing activity, stability, expression, and more in a single round to reach target product profiles faster.
• Adaptive custom model training: After each experimental round, Cradle automatically builds a project-specific model trained on public datasets, proprietary wet lab data, and the customer's own results, ensuring accurate predictions even as designs drift far from natural sequences.
• Intelligent plate design: The system runs tens of thousands of virtual simulations to design curated experimental plates that balance exploiting known winners with exploring new possibilities, minimizing the number of rounds needed to reach a final candidate.
• Hit identification: Analyze high-throughput screening and panning data to rapidly identify promising candidates from large libraries.

Data and Workflow Management

• Flexible data import: Upload wet lab data from screening or optimization experiments using custom schemas that match existing lab workflows, or define goals and assays from scratch for new projects.
• Automated data quality analysis: Cradle automatically assesses experimental data quality and AI learnability, helping teams understand how to refine workflows and assays for faster convergence on optimal sequences.
• Complete provenance tracking: Full audit trails record who did what, when, and with which data — covering every table, task, and configuration across projects.
• Multi-modality support: Engineer enzymes, peptides, and antibodies within a single unified platform via both UI and API.
• Full API access: Connect Cradle directly to existing computational pipelines and automate protein engineering workflows programmatically, without DevOps overhead.

Design and Optimization Features

• Modular task system: Run multiple training and generation configurations side-by-side within each round to compare approaches and identify the best path forward.
• Advanced sequence design controls: Steer designs using template sequences, blocked mutations, regional constraints, and custom optimization objectives tailored to specific project needs.
• Automated model finetuning: Fit-for-purpose models are built automatically for each round using a combination of public data, Cradle's proprietary wet lab data, and company- and project-specific data — all kept strictly private to the customer.
• De-risked plate generation: Models trained on experimental data accurately predict performance in uncharted sequence space, enabling confident engineering of proteins far from wild-type.

Collaboration and Workspace

• Unified team workspace: Wet lab scientists and computational teams share the same data, models, and learning loop, allowing results to compound across the organization.
• Unlimited seats: All team members can participate without per-seat restrictions, encouraging broad adoption across R&D functions.
• Granular access controls: Role-based permissions ensure the right team members access the right projects, data, and reports.
• Web interface and API: Wet lab scientists can generate and review candidates in a few clicks, while computational teams can scale their work programmatically via the API.

Security and Compliance

• Enterprise-grade security: SSO and two-factor authentication (2FA) support protect company intellectual property across the team.
• Data privacy: Models trained on customer data remain strictly private to that customer, with full transparency available via Cradle's security and privacy documentation.

Cradle is built and continuously improved in Amsterdam and Zurich, with its wet lab in Amsterdam providing ongoing experimental validation of every platform feature. The platform integrates into existing experimental workflows without requiring changes to how teams run their labs, delivering upgraded results on the same familiar process.