Abridge vs Nabla (2026)

Abridge is embedded across Epic Haiku, Canto, Hyperdrive and the ASAP emergency module as the first partner in Epic's Partners and Pals and Workshop programs, with bidirectional note sync and write back into the chart. It also has an athenahealth partnership, and Availity's FHIR native payer APIs are embedded for prior authorization.

Sourceabridge.com/press-release/abridge-inside-emoryread 2026-09-29

Patient history and clinical context are retrieved from the EHR, and Nabla writes back to it. Structured notes export into EHR templates, and vital signs, diagnoses and patient instructions are embedded into charts, flowsheets and clinical fields. This works with Epic, athenahealth, Oracle Health, NextGen, Greenway, Arya and ten more EHRs, plus Nabla Connect as an iFrame inside the EHR. Altera Digital Health is also supported. Nabla Connect goes live in days with little engineering, and Nabla runs in 130+ health organizations.

SourceNabla, nabla.com/ehr and nabla.comread 2026-10-05

Fixed function pipelines chain capture, note generation, structured orders, coding context for revenue cycle, and prior authorization with payer determination during the visit. Abridge sets that sequence, and customers do not configure it or combine several agents into a workflow of their own.

Sourceabridge.com/press-release/abridge-availity-collaboration-announcementread 2026-09-29

Ambient notes, dictation and coding run as fixed function products from capture to EHR export. The customer configures no flow, and Nabla names no multiple agents. A generated note can be edited by giving it instructions.

SourceNabla, nabla.com and docs.nabla.com llms.txt indexread 2026-10-05

Ambient capture is started by the clinician across iOS, Android and Epic desktop and mobile surfaces in outpatient, inpatient and emergency settings, including from the ED track board in Haiku, and speech models detect specialty, language and multiple speakers without manual adjustment. Capture happens inside the clinician's workflow, and nothing starts it on a schedule or from an outside event.

Sourceabridge.com/press-release/abridge-inside-for-emergency-medicine-announcementread 2026-09-29

Capture and dictation are started by the clinician inside the EHR or the Nabla app, and Nabla does not say that anything starts them automatically from an event or a schedule. The Connect API creates encounters and their URLs for the EHR, and webhooks report events back to the customer's systems.

SourceNabla, nabla.com and docs.nabla.com llms.txt indexread 2026-10-05

The Contextual Reasoning Engine integrates data from previous patient encounters, health system-specific guidelines and clinician preferences, context-aware evidence surfaces clinical insights from UpToDate, NEJM and JAMA contextualized to the encounter, and payer coverage requirements arrive through Availity's FHIR-native APIs for prior authorization. That context is gathered for each encounter and the literature is chosen by Abridge, and no store of the health system's own documents that the product searches is described.

Sourceabridge.com/abridge-contextual-reasoning-engineread 2026-09-29

Notes draw on patient history and clinical context retrieved from the EHR, and coding suggestions are grounded in the encounter, available EHR context and AMA guidelines. This context is assembled for each encounter, and Nabla names no retrieval structure it maintains. The Core API also extracts ICD-10 and LOINC codes into a FHIR compliant structure.

SourceNabla, nabla.com/coding and docs.nabla.com llms.txt indexread 2026-10-05

Editing tools adapt to clinician preferences, and the Contextual Reasoning Engine carries clinician preferences and data from previous patient encounters into each note. The personalization follows each clinician's style, and how long it lasts, or how a clinician reviews or deletes it, is not described.

Sourceabridge.com/platform/cliniciansread 2026-09-29

Custom instructions saved on a note template and personal dot phrases are written by the clinician and applied by the product on each note. Nabla names no memory layer and no learning of a clinician's style. The API also manages dot phrases and custom dictionary entries, and by default Nabla stores no audio.

SourceNabla, docs.nabla.com llms.txt index and trust.nabla.com; help.nabla.com/en/articles/922050read 2026-10-05

Clinician review and sign off is a required step in the core workflow, since no one can commit a drafted note to the record without a clinician editing and signing it. Linked Evidence exists so the clinician can check each sentence against the source audio before signing.

Sourceabridge.com/press-release/abridge-inside-emoryread 2026-09-29

Coding keeps every final decision in the clinician's hands. Each suggestion arrives with its rationale and is confirmed by the clinician before it is finalized in the record. Nabla shows the reasoning behind each code, tied to the documented clinical evidence.

SourceNabla, nabla.com/codingread 2026-10-05

The trust center lists SOC 2 Type 1 and Type 2 reports validated by an independent auditor for security and confidentiality, plus HIPAA, TX-RAMP and CCPA. Customer facing controls include audit logging, event and audit log management, single sign on, multi factor authentication, role based access, application and network penetration testing, and certificates of destruction, and a security whitepaper, network diagram, model cards and subprocessor list are published.

Sourcetrust.abridge.comread 2026-08-30

Compliance covers SOC 2 Type II, ISO/IEC 27001, HIPAA, GDPR, Texas RAMP Level 2 and CIS Level 2, and note retention is configurable. An Admin Portal lets customer administrators add and delete users. Data is encrypted with TLS 1.2 or higher in transit and AES256 at rest. Nabla also claims compliance with CCPA, PIPEDA, NIST standards and the EU AI Act. It names no single sign on and no user roles.

SourceNabla, trust.nabla.comread 2026-10-05

Linked Evidence makes notes auditable by tying each AI-generated output to its source information across all input data, so every sentence of the note carries a record of what it was drawn from, and the trust center documents detailed audit logs of system activity, including access to and modification of sensitive data, kept in immutable storage. No run or step view of the scribe's work is described.

Sourceabridge-contextual-reasoning-engine and trust.abridge.com; abridge.com/abridge-contextual-reasoning-engineread 2026-09-29

Each coding suggestion shows its reasoning, with the rationale tied to the clinical record. Webhooks from the Connect and Core APIs report events to the customer's systems. Nabla does not mention an audit log or a lasting record of what the scribe read and wrote.

SourceNabla, nabla.com/coding and docs.nabla.com llms.txt indexread 2026-10-05

Abridge runs only as SaaS, hosted on Google Cloud with US based data centers, and offers no customer selectable region, no VPC, no on premises option and no deployment inside the customer's own tenant. It holds TX-RAMP authorization and keeps alternate processing sites for business continuity and disaster recovery.

Sourcetrust.abridge.comread 2026-08-30

Data is hosted on Google Cloud in the US Central region for US clients and in Belgium for non US clients, with speech to text on Azure in US regions and note retention configurable by region. The region follows the client's geography. Nabla does not say whether a customer can pick a different one. Clinical notes are kept 14 days by default, a period each client can change for its region, and audio is not stored by default.

SourceNabla, trust.nabla.comread 2026-10-05

Abridge ships note structures tuned per setting and specialty across 55 plus specialties and 28 plus languages, with purpose built variants for emergency medicine, inpatient and nursing documentation co developed with Mayo Clinic and Epic. These variants are built by Abridge, and there is no library of templates or agents for customers to install.

Sourceabridge.com/press-release/abridge-inside-for-emergency-medicine-announcementread 2026-09-29

Customizable note templates and published specialty custom instructions for cardiology, psychology, pediatrics, dentistry, surgery and weight management are formats and guidance inside one documentation product. Nabla offers no separate agents or template library to adopt. It sells three products (ambient documentation, dictation and coding), used by 85,000+ clinicians.

SourceNabla, nabla.com; help.nabla.com/en/articles/5763202read 2026-10-05

OpenAI and Google Cloud appear as data service vendors on the trust center's subprocessor list. No second model is named, no routing between models is described, and customers and admins cannot choose a model.

Sourcetrust.abridge.comread 2026-08-30

Speech to text runs on Azure in US regions. Nabla names no language model provider and no way for customers to choose or route models. Dictation is clinical grade speech recognition built into Epic and other major EHRs.

SourceNabla, trust.nabla.com and nabla.comread 2026-10-05

No API, SDK, developer portal or MCP server for Abridge's own platform is published, and integration runs through native Epic embedding and partner connections. An API for licensed developers has been reported by a third party research profile, but Abridge does not describe one itself.

Sourceabridge.com, platform/clinicians, abridge-contextual-reasoning-engine and trust.abridge.comread 2026-09-29

Developers building on Nabla's platform get public API documentation at docs.nabla.com. The Core API transcribes audio, generates notes and patient instructions, extracts ICD-10 and LOINC codes as FHIR and supports dictation, dot phrases and custom dictionaries. The Connect API creates encounters, provisions users and sends event callbacks, with OAuth 2.0 client assertions for backends and dated API versions.

SourceNabla, docs.nabla.com llms.txt index and nabla.comread 2026-10-05

Abridge publishes model cards on its trust center, and an independent peer reviewed study with Kaiser Permanente across 1,306 clinicians, published in NEJM AI, evaluated the product. Customers get no testing, evaluation or regression tools of their own, and clinicians' corrections at sign off are the only feedback path.

Sourcetrust.abridge.comread 2026-08-30

Coding quality is checked by regular audits from AAPC certified coders across specialties and by majority vote scoring with ambiguity detection. Nabla runs this quality practice itself and reports 97% or better E/M agreement with expert coders. It names no evaluation or testing surface for customers.

SourceNabla, nabla.com/codingread 2026-10-05

Operates through ambient audio capture and native EHR embedding with write back through Epic's own interfaces; no agent operates third party software through a browser or computer interface.

Sourceabridge.com/press-release/abridge-inside-emoryread 2026-09-29

Nabla embeds in the EHR as an iFrame and exports through integrations. It names no agent that operates a browser or desktop. Dictation runs inside Epic and other major EHRs.

SourceNabla, nabla.com/ehr and nabla.comread 2026-10-05