Artificial intelligence (AI) is rapidly entering clinical settings in a fragmented and largely
unregulated manner. As of May 2025, at least 377 healthcare systems and providers in the U.S. have
piloted or adopted 70 generative AI tools developed by 49 companies for clinical decision support,
patient communication, documentation, claims processing, and healthcare administration. The majority of
American physicians report using AI technologies in clinical care. Similar trends are observed globally:
48% of clinicians across 109 countries report AI use in their work, and large-scale deployments are
emerging across health systems in China.This adoption reflects the strong performance of modern AI models on largely synthetic benchmarks. Large
language models (LLMs) match or exceed physician performance in diagnostic reasoning, clinical text
summarization, medical question answering, patient communication, and multi-step reasoning tasks.
However, concerns about hallucinations, bias, and reliability remain. Real-world clinical performance
remains poorly understood. More realistic benchmarks are emerging, but systematic evaluation in deployed
settings remains rare.A central limitation is the lack of logging for each instance in which a medical AI model is used.
Without consistent records of how models are used in practice, health systems cannot reliably assess
performance, detect failure modes, or measure clinical impact. Reporting frameworks, including TRIPOD+AI,
STARD-AI, DECIDE-AI, SPIRIT-AI, and CONSORT-AI, focus on model development and evaluation in controlled
settings. They do not address continuous, event-level monitoring of deployed systems, especially
generative AI and agentic workflows that rely on prompting, retrieval, and tool use. Several governance
and evaluation frameworks have also been proposed. However, no broadly adopted standard records each
instance of AI use in clinical care.This gap limits the ability to evaluate medical AI in practice and to govern its use. The FDA has
emphasized that clinicians cannot feasibly oversee all outputs of generative AI systems. Medical AI
therefore requires systematic monitoring. In other domains, centralized logging protocols such as
syslog fill this role by recording each system event in a consistent way across many connected
services. These logs support real-time monitoring, root-cause analysis, and auditing at scale. Medical AI
has no equivalent standard.Here we introduce MedLog, a protocol for event-level logging of medical AI. MedLog specifies a schema for
each model invocation, including inputs, outputs, intermediate artifacts, and, when available, clinical
outcomes and user feedback. It records both single-step interactions and multi-stage agentic workflows by
linking events over time. MedLog covers any AI process that uses health data and can affect patient
outcomes, including AI-human and AI-AI interactions. This includes interactions between models and
patients, clinicians, administrators, and other stakeholders; background services such as batch
inference, autonomous triage, claim routing, and continuous monitoring; and AI-AI exchanges within
agentic workflows and orchestration frameworks.
Examples of medical AI interactions that can be logged under the MedLog protocol, as well as the MedLog records they create.
We evaluate MedLog across four clinical deployments spanning intensive care monitoring, infectious
disease severity prediction, hospital quality reporting, and patient attendance prediction. MedLog makes
model behavior visible in practice. It reveals patterns that offline evaluation does not capture,
including temporal failure modes, workflow-dependent variability, interactions between model outputs and
clinician behavior, and performance degradation during severe weather events. By standardizing how health
systems record AI use and link it to outcomes, MedLog creates a basis for continuous evaluation,
auditing, and improvement of medical AI. As AI becomes embedded in clinical workflows, this
infrastructure will be needed to measure real-world performance, detect failures, and guide deployment at
scale.
Quickstart
Emit your first MedLog record in a few API calls.
The record schema
The nine fields of every MedLog record.
Deployments
Four clinical deployments monitored with MedLog.
API reference
Write-once event endpoints with an interactive playground.
{"docsConfig":{"theme":"mint","$schema":"https://mintlify.com/docs.json","name":"MedLog","colors":{"primary":"#2563EB","light":"#3B82F6","dark":"#1D4ED8"},"favicon":"https://mintcdn.com/ayushnoori/SI1z_uchvBg2GlB9/favicon.svg?fit=max\u0026auto=format\u0026n=SI1z_uchvBg2GlB9\u0026q=85\u0026s=40fe00e777ce3fcad972efc59aa7107a","appearance":{"default":"light","strict":false},"navbar":{"links":[{"href":"https://arxiv.org/abs/2510.04033","label":"Read paper"},{"icon":"github","href":"https://github.com/mims-harvard/medlog","label":""}],"primary":{"type":"button","label":"Get involved","href":"/get-involved"}},"footer":{"links":[{"header":"Affiliation","items":[{"label":"Zitnik Lab, Harvard Medical School","href":"https://zitniklab.hms.harvard.edu"}]}]},"seo":{"metatags":{"canonical":"https://medlogprotocol.ai/"}},"styling":{"latex":false},"navigation":{"pages":[]}},"deployment":{"deploymentMetadataValue":{"subdomain":"ayushnoori","hasAudioNarration":false,"actualSubdomain":"ayushnoori","customDomains":["medlogprotocol.ai"],"gitSource":{"type":"github","owner":"ayushnoori","repo":"medlog-docs","deployBranch":"main","contentDirectory":"","isPrivate":true,"mountPath":"","label":""},"entitlements":{"AI_CHAT":{"status":"DISABLED"}},"buildId":"6aab0440029dba9a8c7da672","clientVersion":"0.0.4167","searchFilterCounts":[{"language":"default","tag":"Specification","count":11},{"language":"default","tag":"API Reference","count":6},{"language":"default","tag":"Documentation","count":6},{"language":"default","tag":"Interoperability","count":6},{"language":"default","tag":"Fields at invocation","count":5},{"language":"default","tag":"Endpoints","count":5},{"language":"default","tag":"Events","count":5},{"language":"default","tag":"Standards","count":4},{"language":"default","tag":"Fields appended later","count":4},{"language":"default","tag":"Overview","count":3}]},"assistantConfig":{"enableAskAiButton":false,"enableStarterQuestions":false,"enableContextualStarterQuestions":false,"enableAskAiSearchBar":true},"basePath":"","hasChatPermissions":false,"hasWebMcpPermissions":false,"toggles":[{"name":"kill-isr","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-rbac-roles","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-mintlify-slack","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"agent-integrations-settings","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"rbac-expansion","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-pr-grouping","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"custom-prefix","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"analytics-v7","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"user-flows-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"no-publish-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-onboarding","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving-live","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"stripTemperatureMiddleware","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"support-platform-pylon-zendesk","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"snapshot-serving","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"instant-rollback","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications-server","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"granular-notifications","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automations-fable","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fixed-ai-action-pricing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"automation-resolve-merge-conflict","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"scoped-publishing","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"mintql","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"page-level-analytics","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-version-history-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"billing-page-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"full-page-search","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workspace-projects","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"assistant-v5","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"fork-pr-previews","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"revalidation-jwt","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-sso-scim","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"hobby-abuse-scan","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"dashboard-inbox","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-signup","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-dual-write","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"workos-enabled","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"approve-review-flow","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"editor-metadata","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"sso-setup-deadline-banner","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"turbopuffer-serve","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"live-agents","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false},{"name":"llms-txt-v2","enabled":true,"variant":{"name":"disabled","enabled":false,"feature_enabled":true},"impressionData":false}],"googleFontsInlined":false},"customCss":[{"filePath":"style.css","content":"#footer a{max-width:none;white-space:normal;overflow:visible;text-overflow:clip}#footer a[href=\"/\"] .nav-logo{display:none}#footer a[href=\"/\"]{pointer-events:none;cursor:default;text-decoration:none;font-weight:400;font-size:0.875rem;line-height:1.25rem;color:rgb(2 6 23 / 0.5)}#footer a[href=\"/\"]::after{content:\"An open science, academic research project.\"}.dark #footer a[href=\"/\"]{color:rgb(255 255 255 / 0.5)}#footer div:has(>a[href*=\"github.com\"]){display:none}#footer .md\\:justify-center{justify-content:flex-end!important}.team-grid{display:grid;gap:1.25rem;grid-template-columns:repeat(2, minmax(0, 1fr))}@media (min-width: 768px){.team-grid{grid-template-columns:repeat(3, minmax(0, 1fr))}}@media (min-width: 1024px){.team-grid{grid-template-columns:repeat(4, minmax(0, 1fr))}}.team-card{display:flex;flex-direction:column;align-items:center;padding:1rem;border-radius:0.75rem;border:1px solid rgb(229 231 235);background:white;text-align:center}.dark .team-card{border-color:rgb(31 41 55);background:rgb(17 24 39 / 0.5)}.team-card img{width:5rem;height:5rem;border-radius:9999px;object-fit:cover;margin-bottom:0.75rem;border:3px solid #E5E7EB;background:#F3F4F6}.dark .team-card img{border-color:#374151;background:#1F2937}.team-card-lead img{border-color:#2563EB}.team-card-second img{border-color:#F59E0B}.team-card-deployment img{border-color:#10B981}.team-card-corresponding img{border-color:#8B5CF6}.team-card-name{font-weight:600;font-size:0.875rem;line-height:1.25rem;margin-bottom:0.25rem}.team-card-affil{font-size:0.75rem;line-height:1rem;color:rgb(107 114 128)}.team-legend{display:flex;flex-wrap:wrap;gap:1rem 1.25rem;margin:0 0 1.5rem}.team-legend-item{display:inline-flex;align-items:center;gap:0.5rem;font-size:0.875rem}.team-legend-swatch{display:inline-block;width:0.875rem;height:0.875rem;border-radius:9999px;background:transparent;border:3px solid currentColor}.medlog-figure{display:block;width:100%;height:auto;margin:1.5rem 0 0.75rem}.figure-caption{font-size:0.875rem;line-height:1.4;color:rgb(107 114 128);background:rgb(249 250 251);border:1px solid rgb(229 231 235);border-radius:0.5rem;padding:0.75rem 1rem;margin-bottom:1.5rem}.dark .figure-caption{color:rgb(156 163 175);background:rgb(17 24 39 / 0.5);border-color:rgb(31 41 55)}.iop{--iop-line: rgb(203 213 225);--iop-border: rgb(229 231 235);--iop-surface: white;--iop-surface-2: rgb(249 250 251);--iop-muted: rgb(107 114 128);--iop-blue: #2563EB;--iop-blue-fg: white;--iop-spine: 1.25rem;margin:1.5rem 0 2rem}.dark .iop{--iop-line: rgb(55 65 81);--iop-border: rgb(31 41 55);--iop-surface: rgb(17 24 39 / 0.5);--iop-surface-2: rgb(17 24 39 / 0.7);--iop-muted: rgb(156 163 175);--iop-blue: #60A5FA;--iop-blue-fg: rgb(17 24 39)}.iop-radio{position:absolute;width:1px;height:1px;margin:-1px;padding:0;overflow:hidden;clip:rect(0 0 0 0);clip-path:inset(50%);white-space:nowrap;border:0}.iop-rail-label{font-size:0.6875rem;font-weight:600;letter-spacing:0.06em;text-transform:uppercase;color:var(--iop-muted);margin:0 0 0.5rem}.iop-rail-label+.iop-rail{margin-bottom:1rem}.iop-rail{display:flex;flex-wrap:wrap;gap:0.5rem}.iop-node{display:inline-flex;align-items:center;cursor:pointer;user-select:none;font-size:0.8125rem;line-height:1.2;font-weight:500;padding:0.4rem 0.75rem;border-radius:9999px;border:1px solid var(--iop-border);background:var(--iop-surface);transition:background-color 0.15s ease, border-color 0.15s ease, color 0.15s ease}.iop-node:hover{border-color:var(--iop-blue);color:var(--iop-blue)}#iop-header:checked~.iop-railwrap label[for=\"iop-header\"],#iop-model-instance:checked~.iop-railwrap label[for=\"iop-model-instance\"],#iop-user-identity:checked~.iop-railwrap label[for=\"iop-user-identity\"],#iop-target-identity:checked~.iop-railwrap label[for=\"iop-target-identity\"],#iop-inputs:checked~.iop-railwrap label[for=\"iop-inputs\"],#iop-internal-artifacts:checked~.iop-railwrap label[for=\"iop-internal-artifacts\"],#iop-outputs:checked~.iop-railwrap label[for=\"iop-outputs\"],#iop-outcomes:checked~.iop-railwrap label[for=\"iop-outcomes\"],#iop-user-feedback:checked~.iop-railwrap label[for=\"iop-user-feedback\"]{background:var(--iop-blue);border-color:var(--iop-blue);color:var(--iop-blue-fg)}#iop-header:focus-visible~.iop-railwrap label[for=\"iop-header\"],#iop-model-instance:focus-visible~.iop-railwrap label[for=\"iop-model-instance\"],#iop-user-identity:focus-visible~.iop-railwrap label[for=\"iop-user-identity\"],#iop-target-identity:focus-visible~.iop-railwrap label[for=\"iop-target-identity\"],#iop-inputs:focus-visible~.iop-railwrap label[for=\"iop-inputs\"],#iop-internal-artifacts:focus-visible~.iop-railwrap label[for=\"iop-internal-artifacts\"],#iop-outputs:focus-visible~.iop-railwrap label[for=\"iop-outputs\"],#iop-outcomes:focus-visible~.iop-railwrap label[for=\"iop-outcomes\"],#iop-user-feedback:focus-visible~.iop-railwrap label[for=\"iop-user-feedback\"]{outline:2px solid var(--iop-blue);outline-offset:2px}.iop-card{display:none}#iop-header:checked~.iop-stage .iop-card[data-field=\"header\"],#iop-model-instance:checked~.iop-stage .iop-card[data-field=\"model-instance\"],#iop-user-identity:checked~.iop-stage .iop-card[data-field=\"user-identity\"],#iop-target-identity:checked~.iop-stage .iop-card[data-field=\"target-identity\"],#iop-inputs:checked~.iop-stage .iop-card[data-field=\"inputs\"],#iop-internal-artifacts:checked~.iop-stage .iop-card[data-field=\"internal-artifacts\"],#iop-outputs:checked~.iop-stage .iop-card[data-field=\"outputs\"],#iop-outcomes:checked~.iop-stage .iop-card[data-field=\"outcomes\"],#iop-user-feedback:checked~.iop-stage .iop-card[data-field=\"user-feedback\"]{display:block;animation:iop-fade 0.18s ease-out}@keyframes iop-fade{from{opacity:0;transform:translateY(2px)}to{opacity:1;transform:none}}@media (prefers-reduced-motion: reduce){.iop-card{animation:none!important}.iop-node{transition:none}}.iop-trunk{position:relative;height:1.125rem}.iop-trunk::before{content:\"\";position:absolute;top:0;bottom:0;left:var(--iop-spine);width:2px;background:var(--iop-line)}.iop-hub{border:1px solid var(--iop-border);border-left:3px solid var(--iop-blue);border-radius:0.625rem;background:var(--iop-surface-2);padding:0.875rem 1rem}.iop-hub-tag{font-size:0.6875rem;font-weight:600;letter-spacing:0.06em;text-transform:uppercase;color:var(--iop-muted)}.iop-hub-title{font-size:1rem;font-weight:600;margin-top:0.125rem}.iop-hub-desc{font-size:0.8125rem;line-height:1.5;color:var(--iop-muted);margin-top:0.25rem}.iop-hub-link{display:inline-block;margin-top:0.5rem;font-size:0.75rem;font-weight:600;color:var(--iop-blue)!important;text-decoration:none!important}.iop-hub-link::after{content:\" \\2192\"}.iop-hub-link:hover{text-decoration:underline!important}.iop-branches{position:relative;padding-left:2.25rem;padding-top:0.875rem}.iop-branches::before{content:\"\";position:absolute;top:0;left:var(--iop-spine);width:2px;height:0.875rem;background:var(--iop-line)}.iop-branch{position:relative;padding-bottom:0.75rem}.iop-branch:last-child{padding-bottom:0}.iop-branch::before{content:\"\";position:absolute;top:0;bottom:0;left:-1rem;width:2px;background:var(--iop-line)}.iop-branch:last-child::before{bottom:auto;height:1.375rem}.iop-branch::after{content:\"\";position:absolute;top:1.375rem;left:-1rem;width:1rem;height:2px;background:var(--iop-c)}.iop-branch[data-std=\"uscdi\"]{--iop-c: #10B981;--iop-ct: #047857;--iop-cb: rgb(16 185 129 / 0.45);--iop-cbg: rgb(16 185 129 / 0.08)}.iop-branch[data-std=\"fhir\"]{--iop-c: #F59E0B;--iop-ct: #B45309;--iop-cb: rgb(245 158 11 / 0.5);--iop-cbg: rgb(245 158 11 / 0.1)}.iop-branch[data-std=\"otel\"]{--iop-c: #8B5CF6;--iop-ct: #6D28D9;--iop-cb: rgb(139 92 246 / 0.45);--iop-cbg: rgb(139 92 246 / 0.08)}.dark .iop-branch[data-std=\"uscdi\"]{--iop-ct: #6EE7B7}.dark .iop-branch[data-std=\"fhir\"]{--iop-ct: #FCD34D}.dark .iop-branch[data-std=\"otel\"]{--iop-ct: #C4B5FD}.iop-panel{border:1px solid var(--iop-border);border-left:3px solid var(--iop-c);border-radius:0.625rem;background:var(--iop-surface);padding:0.875rem}.iop-panel-head{display:flex;align-items:center;gap:0.5rem;font-size:0.6875rem;font-weight:700;letter-spacing:0.06em;text-transform:uppercase;color:var(--iop-ct);margin-bottom:0.625rem}.iop-swatch{display:inline-block;flex:none;width:0.5rem;height:0.5rem;border-radius:9999px;background:var(--iop-c)}.iop-chips{display:flex;flex-wrap:wrap;gap:0.375rem}.iop-chip{display:inline-block;font-family:ui-monospace, SFMono-Regular, \"SF Mono\", Menlo, Consolas, monospace;font-size:0.75rem;line-height:1.35;padding:0.25rem 0.5rem;border-radius:0.375rem;border:1px solid var(--iop-cb);background:var(--iop-cbg);color:var(--iop-ct)!important;text-decoration:none!important;overflow-wrap:anywhere;transition:background-color 0.15s ease, border-color 0.15s ease}a.iop-chip::after{content:\" \\2197\";opacity:0.55}a.iop-chip:hover{border-color:var(--iop-c);background:var(--iop-cbg);text-decoration:underline!important}.iop-chip-medlog{--iop-ct: #1D4ED8;--iop-c: #2563EB;--iop-cb: rgb(37 99 235 / 0.45);--iop-cbg: rgb(37 99 235 / 0.08)}.dark .iop-chip-medlog{--iop-ct: #93C5FD}a.iop-chip-medlog::after{content:\"\"}.iop-chip-none{font-family:inherit;font-style:italic;border-style:dashed;border-color:var(--iop-border);background:transparent;color:var(--iop-muted)!important}.iop-note{font-size:0.8125rem;line-height:1.55;color:var(--iop-muted);margin-top:0.75rem}.iop-note code{font-size:0.75rem;overflow-wrap:anywhere}.iop-legend{display:flex;flex-wrap:wrap;gap:0.5rem 1.25rem;margin-top:1.25rem;padding-top:0.875rem;border-top:1px solid var(--iop-border);font-size:0.75rem;color:var(--iop-muted)}.iop-legend-item{display:inline-flex;align-items:center;gap:0.4rem}.iop-legend-uscdi{--iop-c: #10B981}.iop-legend-fhir{--iop-c: #F59E0B}.iop-legend-otel{--iop-c: #8B5CF6}.iop-legend-medlog{--iop-c: #2563EB}@media (max-width: 639px){.iop-branches{padding-left:1.5rem}.iop-branch::before{left:-0.75rem}.iop-branch::after{left:-0.75rem;width:0.75rem}.iop{--iop-spine: 0.75rem}}"}],"page":{"pagePath":"index","slug":"/","pageMetadata":{"title":"What is MedLog?","sidebarTitle":"What is MedLog?","description":"A protocol for event-level logging of medical AI.","href":"/index"},"description":{"compiledSource":"\"use strict\";\nconst {jsx: _jsx} = arguments[0];\nconst {useMDXComponents: _provideComponents} = arguments[0];\nfunction _createMdxContent(props) {\n const _components = {\n p: \"p\",\n ..._provideComponents(),\n ...props.components\n };\n return _jsx(_components.p, {\n children: \"A protocol for event-level logging of medical AI.\"\n });\n}\nfunction MDXContent(props = {}) {\n const {wrapper: MDXLayout} = {\n ..._provideComponents(),\n ...props.components\n };\n return MDXLayout ? _jsx(MDXLayout, {\n ...props,\n children: _jsx(_createMdxContent, {\n ...props\n })\n }) : _createMdxContent(props);\n}\nreturn {\n default: MDXContent\n};\n","frontmatter":{},"scope":{}},"pageType":"default","mdx":{"mdxSource":{"compiledSource":"\"use strict\";const{Fragment:_Fragment,jsx:_jsx,jsxs:_jsxs}=arguments[0];const{useMDXComponents:_provideComponents}=arguments[0];function _createMdxContent(props){const _components={code:`code`,p:`p`,..._provideComponents(),...props.components},{Card,CardGroup,OptimizedImage}=_components;if(!Card)_missingMdxReference(`Card`,true);if(!CardGroup)_missingMdxReference(`CardGroup`,true);if(!OptimizedImage)_missingMdxReference(`OptimizedImage`,true);return _jsxs(_Fragment,{children:[_jsx(_components.p,{children:`Artificial intelligence (AI) is rapidly entering clinical settings in a fragmented and largely\nunregulated manner. As of May 2025, at least 377 healthcare systems and providers in the U.S. have\npiloted or adopted 70 generative AI tools developed by 49 companies for clinical decision support,\npatient communication, documentation, claims processing, and healthcare administration. The majority of\nAmerican physicians report using AI technologies in clinical care. Similar trends are observed globally:\n48% of clinicians across 109 countries report AI use in their work, and large-scale deployments are\nemerging across health systems in China.`}),`\n`,_jsx(_components.p,{children:`This adoption reflects the strong performance of modern AI models on largely synthetic benchmarks. Large\nlanguage models (LLMs) match or exceed physician performance in diagnostic reasoning, clinical text\nsummarization, medical question answering, patient communication, and multi-step reasoning tasks.\nHowever, concerns about hallucinations, bias, and reliability remain. Real-world clinical performance\nremains poorly understood. More realistic benchmarks are emerging, but systematic evaluation in deployed\nsettings remains rare.`}),`\n`,_jsx(_components.p,{children:`A central limitation is the lack of logging for each instance in which a medical AI model is used.\nWithout consistent records of how models are used in practice, health systems cannot reliably assess\nperformance, detect failure modes, or measure clinical impact. Reporting frameworks, including TRIPOD+AI,\nSTARD-AI, DECIDE-AI, SPIRIT-AI, and CONSORT-AI, focus on model development and evaluation in controlled\nsettings. They do not address continuous, event-level monitoring of deployed systems, especially\ngenerative AI and agentic workflows that rely on prompting, retrieval, and tool use. Several governance\nand evaluation frameworks have also been proposed. However, no broadly adopted standard records each\ninstance of AI use in clinical care.`}),`\n`,_jsxs(_components.p,{children:[`This gap limits the ability to evaluate medical AI in practice and to govern its use. The FDA has\nemphasized that clinicians cannot feasibly oversee all outputs of generative AI systems. Medical AI\ntherefore requires systematic monitoring. In other domains, centralized logging protocols such as\n`,_jsx(_components.code,{children:`syslog`}),` fill this role by recording each system event in a consistent way across many connected\nservices. These logs support real-time monitoring, root-cause analysis, and auditing at scale. Medical AI\nhas no equivalent standard.`]}),`\n`,_jsx(_components.p,{children:`Here we introduce MedLog, a protocol for event-level logging of medical AI. MedLog specifies a schema for\neach model invocation, including inputs, outputs, intermediate artifacts, and, when available, clinical\noutcomes and user feedback. It records both single-step interactions and multi-stage agentic workflows by\nlinking events over time. MedLog covers any AI process that uses health data and can affect patient\noutcomes, including AI-human and AI-AI interactions. This includes interactions between models and\npatients, clinicians, administrators, and other stakeholders; background services such as batch\ninference, autonomous triage, claim routing, and continuous monitoring; and AI-AI exchanges within\nagentic workflows and orchestration frameworks.`}),`\n`,_jsx(OptimizedImage,{className:`medlog-figure`,src:`https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=1f100c206ace27046d32d32ff191b919`,alt:`Examples of medical AI interactions that can be logged under the MedLog protocol, as well as the MedLog records they create.`,\"data-og-width\":`3000`,width:`3000`,\"data-og-height\":`1192`,height:`1192`,\"data-path\":`images/medlog-overview.png`,\"data-optimize\":`true`,\"data-opv\":`3`,srcset:`https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=280\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=87521ad93308ed6247cbfea6baf3405b 280w, https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=560\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=1d3e8cffa0d83c94e4116ad2a9b84c81 560w, https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=840\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=816dffcc110c9a1ea2e2cd2f37f30a3a 840w, https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=1100\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=ac235570b0f950ad03639239f009266b 1100w, https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=1650\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=57d4239de7069cc24967f1957e7d03b6 1650w, https://mintcdn.com/ayushnoori/UVFct0BykRx57HBt/images/medlog-overview.png?w=2500\u0026fit=max\u0026auto=format\u0026n=UVFct0BykRx57HBt\u0026q=85\u0026s=41b457b408e7e6f05ac90c1401b1042e 2500w`}),`\n`,_jsx(`div`,{className:`figure-caption`,children:`Examples of medical AI interactions that can be logged under the MedLog protocol, as well as the MedLog records they create.`}),`\n`,_jsx(_components.p,{children:`We evaluate MedLog across four clinical deployments spanning intensive care monitoring, infectious\ndisease severity prediction, hospital quality reporting, and patient attendance prediction. MedLog makes\nmodel behavior visible in practice. It reveals patterns that offline evaluation does not capture,\nincluding temporal failure modes, workflow-dependent variability, interactions between model outputs and\nclinician behavior, and performance degradation during severe weather events. By standardizing how health\nsystems record AI use and link it to outcomes, MedLog creates a basis for continuous evaluation,\nauditing, and improvement of medical AI. As AI becomes embedded in clinical workflows, this\ninfrastructure will be needed to measure real-world performance, detect failures, and guide deployment at\nscale.`}),`\n`,_jsxs(CardGroup,{cols:2,children:[_jsx(Card,{title:`Quickstart`,icon:`bolt`,href:`/quickstart`,children:_jsx(_components.p,{children:`Emit your first MedLog record in a few API calls.`})}),_jsx(Card,{title:`The record schema`,icon:`table-list`,href:`/specification/overview`,children:_jsx(_components.p,{children:`The nine fields of every MedLog record.`})}),_jsx(Card,{title:`Deployments`,icon:`hospital`,href:`/deployments/overview`,children:_jsx(_components.p,{children:`Four clinical deployments monitored with MedLog.`})}),_jsx(Card,{title:`API reference`,icon:`code`,href:`/api-reference/overview`,children:_jsx(_components.p,{children:`Write-once event endpoints with an interactive playground.`})})]})]})}function MDXContent(props={}){const{wrapper:MDXLayout}={..._provideComponents(),...props.components};return MDXLayout?_jsx(MDXLayout,{...props,children:_jsx(_createMdxContent,{...props})}):_createMdxContent(props)}return{default:MDXContent};function _missingMdxReference(id,component){throw new Error(`Expected `+(component?`component`:`object`)+\" `\"+id+\"` to be defined: you likely forgot to import, pass, or provide it.\")}","frontmatter":{},"scope":{"config":{},"pageMetadata":{"title":"What is MedLog?","sidebarTitle":"What is MedLog?","description":"A protocol for event-level logging of medical AI.","href":"/index"}}},"mdxSourceWithNoJs":null},"mdxExtracts":{"tableOfContents":[],"codeExamples":{}},"jsLinks":[],"subdomain":"ayushnoori","siteName":"MedLog","title":"What is MedLog? - MedLog","navMaps":{"firstHrefInVersion":{},"firstHrefInLanguage":{}},"lastModified":"2026-06-01T16:01:10.775Z","originalFileLocation":"index.mdx"},"headHtml":"\u003cscript data-mint-head>(function(){var d=document.documentElement,p=null;if(!p){try{p=localStorage.getItem(\"isDarkMode\")}catch(e){}if(p==='true')p='dark';else if(p==='false')p='light';if(p!=='light'\u0026\u0026p!=='dark'\u0026\u0026p!=='system')p=\"light\"}var r=p==='system'?(matchMedia(\"(prefers-color-scheme: dark)\").matches?'dark':'light'):p;d.classList.remove(r==='dark'?'light':'dark');d.classList.add(r);d.setAttribute(\"data-theme-preference\",p)})();\u003c/script>\u003cmeta data-mint-head name=\"description\" content=\"A protocol for event-level logging of medical AI.\">\u003cmeta data-mint-head name=\"application-name\" content=\"MedLog\">\u003cmeta data-mint-head name=\"generator\" content=\"Mintlify\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/android-chrome-192x192.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"apple-touch-icon\" type=\"image/png\" sizes=\"180x180\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/apple-touch-icon.png\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/favicon-16x16.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/favicon-32x32.png\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/favicon.ico\" media=\"(prefers-color-scheme: light)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"192x192\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon-dark/android-chrome-192x192.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"16x16\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon-dark/favicon-16x16.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"icon\" type=\"image/png\" sizes=\"32x32\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon-dark/favicon-32x32.png\" media=\"(prefers-color-scheme: dark)\">\u003clink data-mint-head rel=\"shortcut icon\" type=\"image/x-icon\" href=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon-dark/favicon.ico\" media=\"(prefers-color-scheme: dark)\">\u003cmeta data-mint-head property=\"og:title\" content=\"What is MedLog? - MedLog\">\u003cmeta data-mint-head property=\"og:description\" content=\"A protocol for event-level logging of medical AI.\">\u003cmeta data-mint-head property=\"og:site_name\" content=\"MedLog\">\u003cmeta data-mint-head property=\"og:url\" content=\"https://medlogprotocol.ai/\">\u003cmeta data-mint-head property=\"og:type\" content=\"website\">\u003cmeta data-mint-head property=\"og:image\" content=\"https://ayushnoori.mintlify.site/_mintlify/api/og/v2?division=Get+started\u0026amp;title=What+is+MedLog%3F\u0026amp;description=A+protocol+for+event-level+logging+of+medical+AI.\u0026amp;theme=f97803526d18f896a60ffe05\u0026amp;signature=fb72e3a2fcaedacc2302fea387fee6e0c67135592d3f105aa2523ea0487ff1f0\">\u003cmeta data-mint-head property=\"og:image:width\" content=\"1200\">\u003cmeta data-mint-head property=\"og:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"twitter:card\" content=\"summary_large_image\">\u003cmeta data-mint-head name=\"twitter:title\" content=\"What is MedLog? - MedLog\">\u003cmeta data-mint-head name=\"twitter:description\" content=\"A protocol for event-level logging of medical AI.\">\u003cmeta data-mint-head name=\"twitter:image\" content=\"https://ayushnoori.mintlify.site/_mintlify/api/og/v2?division=Get+started\u0026amp;title=What+is+MedLog%3F\u0026amp;description=A+protocol+for+event-level+logging+of+medical+AI.\u0026amp;theme=f97803526d18f896a60ffe05\u0026amp;signature=fb72e3a2fcaedacc2302fea387fee6e0c67135592d3f105aa2523ea0487ff1f0\">\u003cmeta data-mint-head name=\"twitter:image:width\" content=\"1200\">\u003cmeta data-mint-head name=\"twitter:image:height\" content=\"630\">\u003cmeta data-mint-head name=\"msapplication-config\" content=\"/mintlify-assets/_mintlify/favicons/ayushnoori/EzvMBmGWoMpC96b4/_generated/favicon/browserconfig.xml\">\u003cmeta data-mint-head name=\"apple-mobile-web-app-title\" content=\"MedLog\">\u003cmeta data-mint-head name=\"msapplication-TileColor\" content=\"#2563EB\">\u003clink data-mint-head rel=\"canonical\" href=\"https://medlogprotocol.ai/\">\u003clink data-mint-head rel=\"alternate\" type=\"application/xml\" href=\"/sitemap.xml\">\u003clink data-mint-head rel=\"alternate\" type=\"text/markdown\" href=\"/index.md\">\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@graph\":[{\"@type\":\"Organization\",\"@id\":\"https://medlogprotocol.ai/#organization\",\"name\":\"MedLog\",\"url\":\"https://medlogprotocol.ai\"},{\"@type\":\"WebSite\",\"@id\":\"https://medlogprotocol.ai#website\",\"name\":\"MedLog\",\"url\":\"https://medlogprotocol.ai\",\"publisher\":{\"@id\":\"https://medlogprotocol.ai/#organization\"}},{\"@type\":\"WebPage\",\"@id\":\"https://medlogprotocol.ai/#webpage\",\"url\":\"https://medlogprotocol.ai/\",\"name\":\"What is MedLog?\",\"description\":\"A protocol for event-level logging of medical AI.\",\"dateModified\":\"2026-06-01T16:01:10.775Z\",\"isPartOf\":{\"@id\":\"https://medlogprotocol.ai#website\"}},{\"@type\":[\"Article\",\"TechArticle\"],\"@id\":\"https://medlogprotocol.ai/#article\",\"headline\":\"What is MedLog?\",\"name\":\"What is MedLog?\",\"description\":\"A protocol for event-level logging of medical AI.\",\"url\":\"https://medlogprotocol.ai/\",\"mainEntityOfPage\":{\"@id\":\"https://medlogprotocol.ai/#webpage\"},\"dateModified\":\"2026-06-01T16:01:10.775Z\",\"publisher\":{\"@id\":\"https://medlogprotocol.ai/#organization\"},\"isPartOf\":{\"@id\":\"https://medlogprotocol.ai#website\"}}]}\u003c/script>\u003cscript data-mint-head type=\"application/ld+json\">{\"@context\":\"https://schema.org\",\"@type\":\"WebSite\",\"name\":\"MedLog\",\"publisher\":{\"@type\":\"Organization\",\"name\":\"MedLog\"},\"provider\":{\"@type\":\"Organization\",\"name\":\"Mintlify\",\"url\":\"https://mintlify.com\"}}\u003c/script>","agentPointerHref":"/index.md","siteName":"MedLog","nav":{"global":null,"tabs":[{"tab":"Documentation","groups":[{"group":"Get started","pages":[{"title":"What is MedLog?","sidebarTitle":"What is MedLog?","href":""},{"title":"Quickstart","description":"Emit your first MedLog record by writing a stream of events to a collector.","href":"/quickstart"}]},{"group":"Deployments","pages":[{"title":"Deployments","sidebarTitle":"Overview","href":"/deployments/overview"}]},{"group":"Implementation","pages":[{"title":"Integration patterns","sidebarTitle":"Integration","href":"/implementation/integration"},{"title":"Storage and retention","sidebarTitle":"Storage and retention","href":"/implementation/storage-retention"},{"title":"Low-resource settings","sidebarTitle":"Low-resource settings","href":"/implementation/low-resource"}]}]},{"tab":"Specification","pages":[{"title":"Specification","href":"/specification/overview"}]},{"tab":"API Reference","pages":[{"title":"API Reference","href":"/api-reference/overview"}]},{"tab":"Interoperability","pages":[{"title":"Interoperability","href":"/interoperability/overview"}]},{"tab":"Team","pages":[{"title":"Team","href":"/team"}]},{"tab":"Citation","pages":[{"title":"Citation","href":"/citation"}]}]},"basePath":"","title":"What is MedLog? - MedLog","buildId":"dpl_9RFrWkXdFTJCytfx6jAofvAaEXr6","deploymentId":"dpl_9RFrWkXdFTJCytfx6jAofvAaEXr6","dynamics":["a74be98ba7"],"chunks":["zoom-image-lightbox-CkI5al_8.js","lightbox-handle-r17bVUz3.js"]}