Asha vs Doximity Ask (formerly DoxGPT): which medical AI fits your workflow?
Doximity has unmatched distribution in US medicine and offers a free, HIPAA-compliant AI assistant (Ask, formerly DoxGPT) plus an ambient note-taker (Scribe). Both run on GPT-4o under the hood. Asha is a fiduciary medical AI with a transparent, citation-grounded corpus, sacred refusals, and an open A2A API. Both tools serve different needs.
TL;DR
Doximity Ask (formerly DoxGPT) is free for verified US clinicians (MDs, DOs, NPs, PAs, pharmacists, podiatrists, CRNAs, and students), HIPAA-compliant, BAA-covered with OpenAI, and primarily designed as an administrative assistant: drafting patient letters, insurance appeals, education materials, chart note templates, and translations. It uses GPT-4o as its backbone; Doximity's own documentation notes the model has limited knowledge of clinical events after October 2023. Doximity Scribe is the ambient note-taker, also free for verified clinicians. Asha is positioned differently: it is a citation-grounded medical AI for both clinicians and patients, with a transparent corpus you can audit (PubMed, OpenAlex, PMC, StatPearls, DailyMed, FDA, clinical guidelines), sacred medical refusals enforced architecturally, and an open A2A API any developer can integrate.
Side-by-side comparison
| Feature | Asha|AI | Doximity Ask (DoxGPT) |
|---|---|---|
| Target audience | Patients, caregivers, clinicians, students, researchers, developers | Verified US clinicians only |
| Primary use case | Evidence-grounded clinical Q&A | Administrative drafting (letters, appeals, patient education, translations) and chart-note support |
| Backbone LLM | Vertex AI Gemini 3.x (primary) + Claude Sonnet (fallback) | GPT-4o (with healthcare RLHF) |
| Citation transparency | Every answer carries verifiable citations to the corpus | Cite Evidence mode references guidelines; the broader DoxGPT product's grounding posture is less explicitly documented |
| Evidence corpus | PubMed (5.1M+), OpenAlex (16.5M top-cited), PMC fulltext, StatPearls, DailyMed, FDA, clinical guidelines, ~126M total vectors | Powered by peer-reviewed content via Doximity's physician network and curated sources (proprietary) |
| Knowledge cutoff | Rolling ingestion; corpus updates continuously | GPT-4o cutoff October 2023 (per Doximity's own published FAQ) |
| Sacred medical refusals | Architecturally enforced (prescriptions, dosing, diagnosis) | Clinician-context tool; refusal posture is editorial |
| HIPAA posture | PHI anonymization, BAA-covered Vertex AI, SOC2 in progress | HIPAA-compliant, BAA in place with OpenAI, SOC2 Type 2 |
| Ambient note-taking (scribe) | Not currently in scope (Arohi covers practice mgmt) | Doximity Scribe (free, 140-min recording, GPT-4o summarization) |
| API access for developers | A2A protocol, self-serve signup | Closed platform; no public developer API |
| Fleet breadth | Asha, Arohi, Lyra, Sage (4 healthcare agents under one architecture) | DoxGPT/Ask + Scribe + Photon e-prescribing + Aledade EHR overlay |
| Distribution | Open web (patient + clinician), A2A API | ~80% of US physicians on Doximity platform broadly; 250K+ prescribers in AI suite |
| Pricing | Free (100/mo); Pro $14.99/mo or $9.99/mo annual; 30-day free trial | Free for verified US clinicians |
When to choose Doximity Ask
Doximity has earned a massive footprint among US clinicians. Choose Doximity Ask when:
- You are a verified US clinician and your primary AI need is administrative: patient letters, insurance appeals, education materials, translations, and chart-note templates.
- You already use Doximity for fax, profile, and dialer features and want one platform.
- You need ambient note-taking and Scribe fits your workflow (HIPAA-compliant, BAA-covered, 140-minute recording).
- You want a tool that integrates with Photon for e-prescribing or Aledade for value-based-care EHR overlay.
- Free is a hard requirement and you accept the GPT-4o knowledge cutoff (October 2023 per Doximity's docs).
When to choose Asha
Asha is built for citation-grounded clinical Q&A with a fiduciary architecture and open developer access. Choose Asha when:
- You want every answer to carry verifiable citations to primary medical literature.
- You serve both clinicians and patients with the same architecture.
- You need an open API to embed medical AI in your own product (A2A protocol, self-serve signup).
- You want a transparent, auditable corpus rather than a proprietary curated set.
- You want sacred refusals on prescription, dosing, and definitive diagnosis built into the architecture.
- You need a knowledge base that updates continuously rather than carrying a 2023-era cutoff.
Specific use cases
Drafting a patient education letter
Doximity Ask: Built for this. Fast, HIPAA-compliant, integrates with Doximity fax line.
Asha: Possible but not the focus. Asha's strength is the medical reasoning behind the letter, not the drafting workflow.
Citation-grounded clinical question (mid-shift)
Doximity Ask: Functional. Cite Evidence mode references guidelines. The underlying model (GPT-4o) has a 2023 cutoff per Doximity's own docs.
Asha: Built for this. Continuous PubMed ingestion, citations on every answer, sacred refusals on prescriptions.
Ambient note-taking during a patient visit
Doximity Scribe: Strong fit. 140-min recording, HIPAA-compliant, EHR-ready output.
Asha: Out of scope. The DNAi fleet covers practice management via Arohi (separate agent).
Patient asking about their own diagnosis
Doximity Ask: Not the audience; requires verified clinician status.
Asha: Built for both clinicians and patients. Same corpus, role-appropriate language.
Pricing breakdown
Doximity Ask and Scribe are free for verified US clinicians (MDs, DOs, NPs, PAs, pharmacists, podiatrists, CRNAs, and students), funded by Doximity's broader platform business (recruitment, advertising, enterprise contracts). Asha is free for 100 queries/month or Asha Pro at $14.99/month ($9.99/month billed annually) for consumers, with A2A API tiers from $49/mo to $999/mo for developers and enterprises. The trade: Doximity is free if you are a verified clinician, Asha is free for everyone with usage caps and offers an API path for developers Doximity does not.
Frequently asked questions
What is the difference between DoxGPT, Doximity Ask, and Doximity Scribe?
Doximity renamed DoxGPT to Doximity Ask. Ask is the chat/search AI assistant (administrative + clinical reference). Scribe is the ambient note-taker that records and summarizes patient visits. Both are free for verified US clinicians.
Is Doximity Ask citation-grounded?
Doximity Ask includes a Cite Evidence mode that references guidelines, and the product is "powered by peer-reviewed content from the Doximity physician network" per the public landing page. The detailed grounding architecture is less explicitly documented than Asha's transparent corpus. As of mid-2026, public reporting on Doximity's grounding methodology is limited; verify with Doximity directly for institutional use cases.
Can I use Doximity Ask if I am not a US clinician?
No. As of mid-2026, Ask and Scribe require verified US clinician status (MD, DO, NP, PA, pharmacist, podiatrist, CRNA, or applicable student/trainee). Asha is open to anyone, including international users and non-clinicians.
Does Doximity Ask have an API?
Doximity's AI integration story is platform-centric (Photon e-prescribing, Aledade EHR overlay, partnerships with health systems). There is no public, self-serve developer API as of mid-2026. Asha's A2A API is self-serve at app.askasha.org/api.
Which one hallucinates less?
Both products use retrieval and guardrails to reduce hallucination versus raw GPT-4o (Doximity) or raw Gemini (Asha). Asha's argument is that the citation-grounded architecture (every answer must cite the corpus) makes hallucinations easier to catch because they are visible in the citations. Independent benchmarking comparing the two specifically is not publicly available; verify with your own test queries.