Sign Up
Login
Try for Free
Login
Freed (often searched as Freed AI) and BastionGPT both offer HIPAA-compliant AI scribes that turn patient visits into structured clinical notes about a minute after the encounter ends. The overlap largely ends there. Freed is an ambient scribe built around the visit: it records the encounter, writes the note, transcribes in 90+ advertised languages, and, on its $119 per month Premier plan, pushes the note into browser-based EHRs and suggests ICD-10 codes. BastionGPT pairs an unlimited AI scribe with a full healthcare AI assistant: it recognizes up to 10 speakers and labels each speaker's name and role automatically, drafts referral letters and insurance appeals, reads 10+ documents at once with 1,000+ pages of context, and runs on named frontier models (GPT-5.5, Claude, and Gemini 3 Pro), starting at $20 per user per month with a signed BAA on every plan.
This comparison covers the differences that matter when you choose: what each platform actually costs in 2026, what happens beyond the visit note, how document handling compares, what each vendor's terms say about your data, and what compliance coverage you get at each price. Freed's pricing, plan limits, and Terms of Use were verified against getfreed.ai and help.getfreed.ai on July 3, 2026.
The quick answer: Choose BastionGPT if you need anything beyond the visit note: analysis of 10+ documents and 1,000+ pages at once, referral letters and insurance appeals, a 10-speaker scribe that labels who said what, and support for 42 CFR Part 2 and FERPA records, with a signed BAA on every plan from $20 per user per month. Freed fits solo clinicians whose documentation needs stop at the visit note, who want one-click note push into a browser-based EHR (a feature of Freed's $119 per month Premier plan), or who regularly document in languages beyond BastionGPT's 15+ (Freed advertises 90+). One hard line to know: Freed's own Terms of Use prohibit processing 42 CFR Part 2 substance use disorder records on its platform, while BastionGPT supports 42 CFR Part 2 alongside HIPAA and FERPA.
| Feature | BastionGPT | Freed |
|---|---|---|
| Entry paid plan | $20/user/month (Professional) | $39/user/month (Starter; capped at 40 notes per month) |
| Full-featured individual plan | $45/user/month (Professional Plus) | $119/user/month (Premier; $104 billed annually) |
| Unlimited scribe notes | Every plan, starting at $20 | Starts at Core, $79/user/month; Starter is capped at 40 notes |
| Free option | 7-day free full trial, plus a 45-day money-back guarantee | 7-day free trial; no permanent free tier or published money-back guarantee |
| Signed BAA | ✓ Every plan, including the free trial | Template BAA embedded in the click-through Terms of Use; signed copy on request; organization-wide BAA listed as a Groups plan feature |
| HIPAA compliant | ✓ | ✓ |
| Additional compliance | 42 CFR Part 2, FERPA, PIPEDA, PHIPA, APP | HITECH, SOC 2 Type II; Terms of Use prohibit processing 42 CFR Part 2 data |
| AI models | GPT-5.5, Claude, Gemini 3 Pro (named and selectable) | Not disclosed |
| Full AI assistant | Unlimited general healthcare assistant on every plan: referral letters, appeals, patient education, document review | Visit-anchored clinician assistant; note editing from Core, medical knowledge base from Premier |
| Document upload and analysis | 10+ documents at once, 1,000+ pages of combined context (Professional Plus and Ultra); PDF, Word, Excel, CSV, PowerPoint, HTML, text, JPG, PNG | Accepts visit audio, uploaded recordings, and transcripts (files up to 256 MB), plus example notes for templates; no general document analysis |
| Multi-speaker recognition | Up to 10 speakers, with diarization and automatic AI labeling of speaker names and roles | Separates speakers and handles accents; no published speaker count; independent 2026 reviews report limited diarization for multi-party visits |
| Max recording length | 240 minutes per session | No published limit; audio for visits left paused over 14 days is not retained |
| Note types | SOAP, Detailed SOAP, DAP, BIRP, H&P, progress, custom free-form prompt | Specialty templates across 30+ specialties; template builder from Core; learned formats |
| Saved prompts with team sharing | Saved Prompts library with folder organization and team sharing on every plan | Custom templates are per clinician on individual plans; team template sharing is a Groups plan feature |
| Coding support | Medical coding suggestions and under- or over-coding review through the assistant, on every plan | ICD-10 suggestions on Premier only; CPT in beta |
| EHR compatibility | Works alongside 50+ EMR/EHR systems (Epic, Oracle Health, Athenahealth, eClinicalWorks) | Copy-paste on all plans; one-click EHR push to browser-based EHRs on Premier (Athenahealth, Practice Fusion, SimplePractice, Elation, Tebra) |
| AI training on your data | Customer data is never used to train AI models | States identifiable PHI is not used for training; Terms permit model training on customer data with consent via settings, and de-identified data use and disclosure at Freed's discretion |
| Languages | 15+ | 90+ advertised, with code-switching and automatic detection; notes are generated in English |
| Uptime guarantee | 99.90% | No published SLA |
| Support | Email, chat, phone, and video with real humans, on every plan | Chat, email, and live support calls |
| Verified user ratings | 4.8/5 average across Capterra, Software Advice, and GetApp | 4.8/5 on the Apple App Store |
Both platforms offer HIPAA-compliant ambient transcription, generate structured clinical notes within a minute or two of ending a visit, and deploy in minutes without an IT project. Both run on Microsoft Azure infrastructure with US-based data storage, encrypt data in transit and at rest, learn your note format from your edits, and work for in-person and telehealth visits. Both state that identifiable patient data is never used to train AI models. If your entire requirement is turning a spoken visit into a standard note, both tools cover that core workflow. The differences appear the moment you price full access, upload a patient record, run a group session, document substance use disorder treatment, or need anything beyond the visit note.
Freed's current lineup: Starter at $39 per user per month, capped at 40 notes per month; Core at $79 with unlimited notes and an editing assistant; Premier at $119 per month ($104 billed annually) with EHR push, ICD-10 suggestions, visit summaries, and the medical knowledge assistant; and custom Groups pricing. Freed's help center notes that small groups of around four clinicians are billed at the standard individual rate, with custom pricing starting at 10 or more clinicians.
BastionGPT Professional starts at $20 per user per month with the unlimited AI scribe and unlimited AI assistant included. Professional Plus, the closest match to Freed's Premier tier, costs $45 and adds multi-document analysis, image and spreadsheet reading, and manual model selection. Prices have held steady for over two years, and annual billing includes a free month.
The math for a five-clinician practice: Freed Premier runs $7,140 per year at monthly billing ($119 x 5 x 12), or $6,240 billed annually. BastionGPT Professional Plus runs $2,700 ($45 x 5 x 12). That is a difference of $4,440 per year, or $3,540 if Freed is billed annually. Even Freed's mid-tier Core plan, without EHR push or coding, costs $4,740 per year for the same five clinicians: $2,040 more than BastionGPT Professional Plus.
Freed's Clinician Assistant works inside the visit workflow: it edits the note (from the Core plan) and, on Premier, answers clinical questions from a medical knowledge base with linked citations. It is tied to the encounter, and its most useful capabilities sit behind the $119 tier.
BastionGPT includes an unlimited general healthcare AI assistant on every plan. Clinicians use it to draft referral letters, insurance appeal letters, treatment plans, and patient education materials; review documentation for incorrect names, pronouns, vague language, and suspected under- or over-coding; and analyze lab summaries and spreadsheets. BastionGPT earned 99.3% sentence-level accuracy on clinical documentation tasks from a peer-reviewed study by Baylor College of Medicine and Texas Children's Hospital researchers.
This is the widest gap between the two platforms.
On Professional Plus and Ultra plans, BastionGPT reads 10+ documents at once, up to 1,000+ pages of combined context. Supported formats include PDF, Word, Excel, CSV, PowerPoint, HTML, plain text, and JPG or PNG images, including scanned PDFs, charts, and graphs. A clinician can upload a full patient history, a psychological evaluation, and records from multiple providers, then query, summarize, and cross-reference all of it in one conversation. The Professional plan covers a single PDF, Word, or text document up to 30 pages.
Freed's platform accepts visit audio, uploaded recordings, and transcripts (files up to 256 MB), plus example notes for building templates. It does not offer general document upload and analysis: there is no workflow for reading a PDF record, an insurance document, or a spreadsheet. Freed's pre-visit context comes from prior Freed notes for the same patient, not from outside records. If reviewing incoming records is part of your day, that work stays manual with Freed.
BastionGPT's AI Scribe recognizes up to 10 speakers in a single session, live or from uploaded audio, for sessions up to 240 minutes. Speaker diarization separates the transcript by voice, and BastionGPT's AI then labels each speaker's name and role automatically (for example, clinician, patient, or parent) directly in the transcript. Group therapy, family sessions, care conferences, and multi-provider visits produce a transcript that reads like a script instead of a wall of text.
Freed separates speakers, handles accents, and offers an opt-in Voice ID that recognizes the clinician's own voice for attribution. It publishes no speaker count and offers no automatic labeling of speaker names and roles, and independent 2026 reviews report that advanced diarization for group visits and multi-party encounters is limited. For practices that run family or group sessions, that difference shows up in every transcript.
This difference is easy to miss and expensive to discover late. Freed's Platform Terms of Use state that the platform "has not been designed to process Part 2 Data" and require users to agree not to process any 42 CFR Part 2 data, with Freed disclaiming liability if they do. 42 CFR Part 2 covers records from federally assisted substance use disorder programs, a common record type in mental and behavioral health, addiction medicine, and integrated primary care.
BastionGPT supports HIPAA, 42 CFR Part 2, and FERPA, along with PIPEDA and PHIPA (Canada) and APP (Australia). Practices that touch SUD treatment records or school-based health records can use the same platform for all of their documentation instead of carving out exceptions.
BastionGPT names its engines: licensed and HIPAA-compliant versions of GPT-5.5, Claude, and Gemini 3 Pro. Every plan selects the best model for each task automatically, and Professional Plus adds manual model selection plus early access to new models as they ship. Model quality drives note quality, and naming the models lets you verify exactly what is reading your records.
Freed does not disclose which foundation models power its product; third-party reviews describe Azure-backed language models without naming them.
Both vendors state that identifiable patient data is never used to train AI models. The difference is in the rights each vendor reserves. Freed's Terms of Use grant Freed a license to train its AI/ML models on customer data with consent provided via platform settings, give Freed the right to use de-identified data "in our sole discretion" and disclose it to third parties, and permit Freed to sell aggregate data. Retention is user-controlled: keep notes indefinitely or auto-delete after 30 days, with backups persisting seven additional days.
BastionGPT's position is simpler: customer inputs, documents, and audio are never used to train AI models, on any plan, under any setting. Chats and transcripts are wiped after 30 days by default, and users can retain longer or delete sooner. Data is encrypted with AES-256 at rest and TLS in transit, runs on HITRUST CSF certified and SOC 2 Type II attested Microsoft Azure infrastructure, and is independently penetration tested at least annually. In three years serving more than 10,000 healthcare organizations, BastionGPT has had zero breaches of protected health information.
Beyond the standard formats, BastionGPT's Custom Note tab accepts any free-form prompt per recording: narrative formats, parent-friendly summaries, forensic documentation, and other specialty outputs. You can also upload example notes so BastionGPT matches your writing style and formatting. A Saved Prompts library with folder organization and team sharing is included on every plan, so a practice can standardize documentation by sharing one vetted prompt set across all providers and insert any prompt into a chat with a slash command.
Freed offers a template builder from its Core plan and a "learn my format" tool, and its templates cover 30+ specialties. On individual plans those custom templates stay with the clinician who built them; team template sharing is a feature of Freed's custom Groups plans.
BastionGPT enters a signed Business Associate Agreement with every customer on every plan, including the free trial. Freed also provides a BAA: it is embedded as a schedule in the click-through Platform Terms of Use accepted at signup, a signed copy is available on request, and an organization-wide BAA is listed as a feature of Freed's custom Groups plans. Practices whose compliance policies require a countersigned BAA in hand before the first recording should factor in that request step; with BastionGPT the signed agreement is part of onboarding on every plan.
BastionGPT publishes a 99.90% uptime guarantee and offers email, chat, phone, and video support with real humans on every plan. Freed offers US-based support through chat, email, and live support calls, and maintains a public status page; it does not publish an uptime SLA. Recent Apple App Store reviews praise Freed's accuracy and time savings, but also reports lost notes after long visits and app access issues.
Freed's clearest workflow advantage is its one-click EHR push: on the Premier plan, a Chrome extension maps the finished note into the fields of browser-based EHRs, with Athenahealth, Practice Fusion, SimplePractice, Elation, and Tebra confirmed. BastionGPT works alongside 50+ EMR/EHR systems through copy-paste and document uploads, which covers desktop and browser EHRs alike but adds a paste step Freed's Premier users skip. Freed also embeds ICD-10 suggestions (with CPT in beta) directly in the scribe workflow on Premier, and its pre-visit prep surfaces summaries of the patient's prior Freed notes before the encounter. Both are genuine workflow conveniences for clinicians who live entirely inside the visit-to-note loop. Freed's strongest published number is language coverage: it advertises transcription in 90+ languages with code-switching and automatic detection, so a visit can move between languages mid-conversation, and the finished note is generated in English. BastionGPT supports 15+ languages. If your practice regularly documents visits in languages BastionGPT does not yet cover, that difference can be significant. Weigh those strengths against what they cost. The plan that includes EHR push and coding runs $119 per user per month, the platform accepts no outside documents, its terms exclude 42 CFR Part 2 records, and the assistant stays anchored to the visit. For US practices that document primarily in English and need records review, appeal letters, group session transcripts, or spreadsheet analysis, BastionGPT delivers more capability at less than half of Freed's full-access price.
"BastionGPT has a lot more capabilities [than ChatGPT] with data protection, so clinicians can more freely use AI."
"People compliment the quality and clarity of my documentation all the time now. I can't imagine working without Bastion!"
"BastionGPT has reshaped the way I practice medicine, teach, and manage day-to-day operations."
Full access to the AI scribe and AI assistant. You won't be charged today, you can cancel anytime, and every plan includes a 45-day money-back guarantee.
Freed's current plans are: Starter at $39 per user per month (capped at 40 notes per month), Core at $79 (unlimited notes), Premier at $119 per month or $104 billed annually (adds EHR push, ICD-10 suggestions, visit summaries, and the medical knowledge assistant), and custom Groups pricing. Small groups of around four clinicians pay standard individual rates, per Freed's help center. BastionGPT plans run $20 to $75 per user per month.
Freed offers a 7-day free trial with no credit card required, and a 50% student discount on the monthly individual plan. There is no permanent free tier and no published money-back guarantee. BastionGPT offers a 7-day free trial and plans from $20 per user per month with a 45-day money-back guarantee.
Freed provides a Business Associate Agreement embedded as a schedule in its click-through Platform Terms of Use, accepted automatically at signup; a signed copy is available on request, and an organization-wide BAA is listed as a Groups plan feature. BastionGPT includes a signed BAA on every plan, including the free trial.
No. Freed's Platform Terms of Use state the platform has not been designed to process Part 2 data and require users to agree not to process any 42 CFR Part 2 data, with liability disclaimed if they do. BastionGPT supports 42 CFR Part 2 alongside HIPAA, FERPA, PIPEDA, PHIPA, and APP.
No. Freed accepts visit audio, uploaded recordings, and transcripts (files up to 256 MB), plus example notes for templates. It does not offer general document upload and analysis of PDFs, records, images, or spreadsheets. BastionGPT reads 10+ documents and 1,000+ pages at once on Professional Plus and Ultra, including Excel and CSV files.
BastionGPT reads Excel (XLS, XLSX, XLSM), CSV, PowerPoint, and HTML files on Professional Plus and Ultra, alongside PDF, Word, text, and image files. Freed's published upload support covers visit audio, video, and text transcripts for note generation; it does not list spreadsheet upload or analysis.
On Professional Plus and Ultra, BastionGPT reads 10+ documents at once with 1,000+ pages of combined context. The Professional plan covers a single PDF, Word, or text document up to 30 pages. Supported formats include PDF, Word, Excel, CSV, PowerPoint, HTML, text, and images.
BastionGPT recognizes up to 10 speakers per session, live or from uploaded audio, for sessions up to 240 minutes, with diarization and automatic AI labeling of each speaker's name and role. Freed separates speakers and offers an opt-in Voice ID for the clinician's voice; it publishes no speaker count, and independent 2026 reviews report limited diarization for multi-party visits.
BastionGPT uses licensed versions of GPT-5.5, Claude, and Gemini 3 Pro, selecting the best model for each task automatically; Professional Plus adds manual model selection and early access to new models. Freed does not disclose which foundation models power its product.
Freed states identifiable PHI is not used for AI training. Its Terms of Use permit training its AI/ML models on customer data with consent provided via platform settings, allow Freed to use and disclose de-identified data at its sole discretion, and permit sale of aggregate data. BastionGPT never uses customer inputs, documents, or audio to train AI models, on any plan.
Yes. Freed advertises transcription in 90+ languages with code-switching and automatic detection, and generates the finished note in English. BastionGPT supports 15+ languages. If extensive multilingual transcription is your deciding requirement, weigh it against the differences in document analysis, assistant scope, 42 CFR Part 2 coverage, and price.
Freed Premier includes one-click EHR push into browser-based EHRs (Athenahealth, Practice Fusion, SimplePractice, Elation, and Tebra are confirmed), with copy-paste on all plans. BastionGPT works alongside 50+ EMR/EHR systems, including Epic, Oracle Health, Athenahealth, and eClinicalWorks, through copy-paste and document uploads. Choose Freed's Premier tier if one-click push into a supported browser EHR is your deciding requirement.
BastionGPT is the stronger fit for most mental and behavioral health practices: it supports 42 CFR Part 2 records (Freed's terms prohibit them), transcribes group and family sessions with up to 10 labeled speakers, produces DAP and BIRP notes with custom free-form prompts, and analyzes prior records and psychological evaluations. Freed remains a capable scribe for solo therapists whose documentation stops at the individual session note and who avoid Part 2 data.
BastionGPT Professional Plus costs $2,700 per year for five clinicians ($45 x 5 x 12). Freed Premier costs $7,140 at monthly billing or $6,240 billed annually, and Freed's help center notes small groups of around four clinicians pay standard individual rates. That is a difference of $3,540 to $4,440 per year, and BastionGPT includes document analysis and the full assistant at that price.
See how BastionGPT stacks up against other AI tools for healthcare:
