27 tools from the Useful AI MCP Server, categorised by risk level.
View the Useful AI policy →suggest Suggest a tool that should exist on the Useful AI platform but doesn't yet. useful Useful AI has 183+ tools beyond the named ones above. Describe any task involving data transformation, code generation, text analysis, validation, ... useful_address_string_parser Batch-parse freeform addresses into structured components with validation, abbreviation expansion, and canonical formatting. useful_batch_record_normalizer Normalize mixed-format vendor records into canonical IDs, integer USD thresholds, and boolean eligibility in one batch call. useful_batch_slug_generator Batch-convert multilingual titles into collision-free URL slugs with ready-to-use SEO meta tags. useful_certificate_chain_inspector Parse X.509 certificates, display all fields, verify chain order, and flag expiring or weak certs. useful_color_contrast_fixer Parse mixed CSS colors, compute WCAG contrast ratios, and output fixed compliant color pairs in one batch call. useful_contact_entry_parser Batch-parse messy international phone numbers and addresses into structured E.164 phones, split address fields, and confidence flags. useful_cross_source_field_reconciler Reconcile fields, units, IDs, and types across mismatched data sources into one executable merge plan. useful_css_inline_tool Inline CSS into HTML elements, expand shorthands, and flag un-inlineable selectors like :hover and @media. useful_dispatch Useful AI has 183+ tools beyond the named ones above. Describe any task involving data transformation, code generation, text analysis, validation, ... useful_food_temperature_inspection_parser Batch-parse messy inspection temp strings into structured F/C values with FDA compliance flags per context. useful_income_narrative_batch_normalizer Batch-parse caseworker income narratives into canonical monthly gross USD integers with component breakdowns and confidence flags. useful_logit_lens_simulator Visualize how LLM predictions evolve layer by layer, revealing where confident errors crystallize in transformers. useful_mock_data_generator Generate hundreds of schema-compliant, internally consistent fake records from a JSON schema, seed, and row count. useful_ndc_batch_normalizer_and_formulary_validator Batch-extract, zero-fill-normalize, and formulary-validate NDCs from mixed-format and composite SKU strings. useful_password_policy_validator Validate passwords against configurable policies with entropy, breach check, rule results, and strength score in one call. useful_phone_number_normalizer Batch-validate, classify, and reformat phone numbers from any messy format into E.164 and local display formats. useful_phone_number_parser Batch-parse messy phone numbers into E.164 format with country detection, validation, and mobile/landline type classification. useful_prompt_drift_detector Detect how subtle prompt edits silently shift model behavior and alignment across iterations. useful_prompt_x_ray AI-powered structural analysis that dissects prompts into layers, finds conflicts, and reveals hidden failure modes. useful_recurrence_rule_expander Batch-parse mixed recurrence formats into canonical rules and expand them into concrete occurrence dates. useful_royalty_escalation_clause_parser Batch-parse free-text royalty escalation clauses into sorted, validated JSON tier arrays with gap/overlap flags. useful_suggest Suggest a tool that should exist on the Useful AI platform but doesn't yet. The platform builds frequently requested tools automatically. useful_token_attention_heatmap Visualize token attention patterns and context fragmentation in your prompts with real semantic analysis. useful_xml_schema_validator Validate XML against XSD schemas, get structured error reports with line numbers and fix suggestions. The Useful AI MCP server exposes 27 tools across 2 categories: Read, Execute.
Use Intercept, the open-source MCP proxy. Write YAML rules for each tool — rate limits, argument validation, or deny rules — then run Intercept in front of the Useful AI server.
Useful AI tools are categorised as Read (26), Execute (1). Each category has a recommended default policy.
Open source. One binary. Zero dependencies.
npx -y @policylayer/intercept