# Medical Record Review API and Integrations > Integrate medical record processing, source-linked chronologies and structured medical case data with your product. Medrecords AI APIs and deployment options. Canonical page: https://medrecords.ai/partners/api/ --- New — **Missing Records Detection:** flags every visit, provider, and date missing from the file. [See how →](https://medrecords.ai/product/missing-records-identification/) [Home](https://medrecords.ai/) ›[Partners](https://medrecords.ai/partners/) › — API and integrations Medical record review API ## Build medical record intelligence into your existing software. Add supported medical record ingestion, structured case analysis, source-linked chronology output and reporting workflows to your product, without building the processing engine. By [Ahmed Jemaa](https://medrecords.ai/authors/ahmed-jemaa/), Co-Founder & CEO of Medrecords AI · Published 12 Oct 2026 · Updated 12 Oct 2026 [Request a technical demo](https://medrecords.ai/partners/api/#apply) [Integration overview](https://medrecords.ai/partners/api/#how) REST cases and events — ·Evidence graph with page anchors — ·Hosted, your cloud, or your network ### Integrate where your users already work. The 6 places a cited chronology changes the product it sits inside. #### Case management Create cases from intake and keep processing state in sync, so the reviewer never leaves the queue. #### Medical record retrieval Add a structured review service after document collection, delivered inside the same portal. #### Claims administration Surface a source-linked medical timeline inside the claims workflow, next to the reserve. #### IME and QME platforms Populate draft report workflows for authorized evaluators from the assembled, cited record. #### Review vendors Push outputs into existing reviewer queues and stop re-keying facts between systems. #### Legal-tech applications Build case-specific search, chronology and report experiences on the evidence graph. ### Structured outputs, not another manual export step. Ingest, process, retrieve. Events and evidence data populate your own UI; nobody downloads a PDF and re-types it. **Your application** Intake, case management, claims or reviewer queue → **Case over REST** Create the case, reference authorized documents → **Processing** Read, sort, deduplicate, extract, cite → **Events and structured data** State changes pushed to you; facts and citations queried → **Your reviewer, your UI** Chronology and report rendered where your users already work Illustrative flow. Endpoint paths, authentication and the event list come from the technical documentation after approval, not from this page. ### What the API covers. Each capability with where it stands today. "Confirm on the call" means it exists in the product and the exact contract for your integration is settled in discovery. | Capability | What it does | Status | | --- | --- | --- | | Case creation | Create and manage cases over REST. Medical Data API → | Available | | File ingestion | Upload or reference authorized documents; size, format and transport confirmed per integration | Confirm on the call | | Processing events | Subscribe to the events you care about instead of polling | Available | | Evidence graph | Query extracted facts and relationships by case, entity type or date range. Evidence Graph API → | Available | | Source citations | Every fact carries its page anchor and source document | Available | | Chronology and report outputs | Structured content or branded exports. Export controls → | Available | | Authentication | Credentials scoped to your tenancy after the agreement | Confirm on the call | | Tenant isolation | Partner and client segregation; architecture review on request | Confirm on the call | | Rate limits and quotas | Agreed for your volume in discovery | Confirm on the call | | Environments | Synthetic-case testing before production; sandbox shape agreed per integration | Confirm on the call | **** [](https://medrecords.ai/product/medical-data-api/) **** **** **** [](https://medrecords.ai/product/evidence-graph-api/) **** **** [](https://medrecords.ai/product/branding-export-controls/) **** **** **** **** ### A practical path from proof of concept to production. 6 steps, in the order they happen. Regulated data does not move until step 3 is signed. 1. **Describe the requirement** Case volume, average pages, and the outputs your product needs. 2. **Technical fit call** Authentication, tenancy, file transfer, events, retention. Your engineers and ours. 3. **Agreement and permissions** The right contract for the data involved, BAA included where PHI is in scope. 4. **Test with synthetic cases** Non-PHI sample cases in the approved environment. No production data yet. 5. **Validate the integration** Data mapping, citation integrity, error behaviour, retries and the reviewer handoff. 6. **Production** Contracted scope, deployment, observation, iteration. ### Deploy for your architecture and compliance needs. Data residency, network access and processing model differ per option; each is documented in its agreement. #### Managed platform and API Hosted by Medrecords AI, integrated over the API. Fastest to a proof of concept. #### Your cloud tenancy A flat annual license with the platform in your own AWS or Azure account and model inference billed there. [Pricing →](https://medrecords.ai/pricing/) #### Your own network The whole stack inside your infrastructure, air-gap available, for the contracts that require it. [Deployment options →](https://medrecords.ai/pricing/) ### The questions your security and engineering reviewers will ask. Bring them to the technical call; where the answer is "it depends on your integration" we say so instead of guessing. | Topic | Where the answer lives | | --- | --- | | PHI, BAA and permitted uses | HIPAA and BAA · DPA | | Encryption and credential handling | Trust Center | | Tenancy, logs and auditability | Architecture review on the technical call | | Uptime and support commitments | Set out in the agreement; no blanket SLA is published here | | Error handling, idempotency, concurrency | Technical documentation after approval | | Versioning and deprecation | Technical documentation after approval | | Webhook security | Technical documentation after approval | | Data retention and deletion | DPA and the agreement | | Model training terms | Customer records never train a model; contractual under the BAA | | Source provenance | Verifiable citations · Evidence graph | **** [](https://medrecords.ai/hipaa/) [](https://medrecords.ai/dpa/) **** [](https://medrecords.ai/security/) **** **** **** **** **** **** [](https://medrecords.ai/dpa/) **** **** [](https://medrecords.ai/product/citations/) [](https://medrecords.ai/product/evidence-graph-api/) ### Pricing for your use case and volume. Usage or enterprise licensing, depending on throughput, hosting, support and integration needs. Tell us your monthly case volume and average pages and we scope a pilot. No per-call fees or SLA tiers are published; what we quote is written into the agreement. Self-Service meters deduplicated pages from 10 cents down to 5 cents at volume, duplicates never billed; the 2 annual licenses are on the [pricing page](https://medrecords.ai/pricing/). API partners usually land on a license. ### API and integrations, answered. ##### Is the API generally available or by approval? By approval during a technical discovery, so that tenancy, data agreements and volume are settled before credentials exist. There is no self-serve key page today, and we would rather say that than imply one. ##### Is there a sandbox? Testing happens with synthetic, non-PHI sample cases in an approved environment before any production data moves. Whether that is a separate sandbox tenancy or a scoped test account for your integration is agreed on the technical call. ##### How are API credentials provisioned? After the agreement, by the Medrecords AI team, scoped to your tenancy. Rotation and revocation are covered in the security documentation. Discuss the specifics with our technical team. ##### What file types are supported? PDFs, scanned images and handwritten pages are read; DICOM imaging has its own review path. Transport, size limits and batch behaviour are confirmed for your integration. [Medical records OCR →](https://medrecords.ai/product/ocr/) ##### What are the processing states and events? A case moves from created through processing to ready, and your endpoint receives events as it does, so your system stays current without polling. The event list is confirmed against the current documentation on the call. [Medical Data API →](https://medrecords.ai/product/medical-data-api/) ##### Are citations returned programmatically? Yes. Every extracted fact carries its page anchor and source document, so citations survive the API boundary into your own system. Field names and resolution are in the documentation. [Evidence Graph API →](https://medrecords.ai/product/evidence-graph-api/) ##### Can it run inside our cloud or network? Yes. There is a flat annual license in your own cloud tenancy and one inside your own network, air-gap available. Responsibilities per option are set out in the agreement. [Pricing and deployment →](https://medrecords.ai/pricing/) ##### How are rate limits and usage billed? Usage is metered on deduplicated pages, or covered by an annual license without metering. Rate limits and throughput for your volume are agreed in the technical discovery, not published as a single number. ##### What are the security and data-processing requirements? A BAA where PHI is involved, a DPA, and the permitted-use terms in the agreement. Encryption, tenancy, logging and retention are documented in the [Trust Center](https://medrecords.ai/security/). Records never train a model. ##### How long does an integration take? A proof of concept with synthetic cases is measured in weeks; production depends on your data mapping, error handling and reviewer handoff. We give a timeline on the technical call, after seeing your architecture, not before. ### Let's design the integration around your product. Tell us what your product does and what it needs from the record. The technical team replies with the next step: a fit call, documentation eligibility, or an honest "not yet". **No patient records, names or case details in this form.** Secure intake instructions come after the first call. Prefer email? [aj@medrecords.ai](mailto:aj@medrecords.ai). #### Thanks. The technical team has it. You will hear from aj@medrecords.ai with a proposed fit call and what to bring to it. If a calendar is faster, [pick a 30-minute slot](https://medrecords.ai/demo/) and mention the API. ### More ways to work with Medrecords AI - [**White-label solutions** — Deliver chronologies and reports under your brand. For review companies, IME networks and BPOs.](https://medrecords.ai/partners/white-label/) - [**Referral program** — Introduce qualified buyers from your network. Application only while the pilot runs.](https://medrecords.ai/partners/referrals/) - [**Published rates for review vendors** — The per-page rate and the 2 flat licenses, so you can price your own work against them.](https://medrecords.ai/solutions/medical-record-review-vendors/)