Medi vs Candid Health

Short answer

Candid Health and Medi solve different problems for different buyers. Candid is an AI-native RCM automation platform built for digital health companies, multi-site provider groups, and managed service organizations that have engineering teams and want billing wired into their product infrastructure. It raised a $52.5 million Series C in February 2025 (about $99.5 million total), prices on a percentage of collections by most third-party accounts, and goes live through an API integration project. Medi is built for independent third-party billing companies: $20 per client practice per month with published graduated volume discounts (1-25 practices at $20, 26-50 at $15, 51+ at $10 each), per-transaction EDI pricing, no per-provider fees, no contract required, and no EHR dependency. Medi publishes its full schedule including volume discounts; Candid is quote-based and does not publish rates. If you run a billing company that manages multiple client practices, the question is not which AI is better. It is whether you need infrastructure you integrate into or software you operate. For most independent billers, that points away from Candid. See candidhealth.com for their current product and the TechCrunch Series C coverage for funding context.

Choose Medi if billing is your business

• You run an independent billing company managing five or more provider practices.
• You want a per-practice monthly cost that does not change when a client practice adds providers, and that falls on a published graduated scale as your book of practices grows.
• You need multi-practice operator workflows: shared work queues, a denial queue, and ERA review across every client practice on one screen.
• Your team is billers, not engineers. You want to log in and work, not build and maintain an API integration.
• You want transparent per-transaction EDI pricing: $0.25 first claim line, $0.20 each additional; $0.25 first paid ERA line, $0.20 each additional, $0 for denied ERA lines after the first; $0.20 eligibility and status; $1 COB, discovery, and attachments.
• You want to avoid monthly cost that moves with collection volume.
• You want software built around the billing company as the primary tenant, not the practice or the provider.

Choose Candid Health if billing is part of your product

• You are a digital health company, telehealth provider, or multi-site medical group that wants billing automation embedded in your product through an API.
• You have an engineering team that can own an integration and maintain it as your stack changes.
• You are a scaling digital health company billing insurance for the first time and want a managed billing-infrastructure partner rather than a software tool you operate.
• Your claim volume is high enough that a percentage of collections lands below a flat-rate alternative.
• You want AI-driven touchless claim processing and are willing to pay a percentage of collections for it.

Pricing model

Candid does not publish a price. The percentage-of-collections figure comes from third-party analysis at Out-of-Pocket Health and varies by volume, specialty, and contract. Request a current quote from Candid directly. Medi publishes its complete rate schedule, including graduated volume discounts, on its pricing page.

Where Candid Health is stronger

Candid's AI rules engine is built for high-volume, first-pass accuracy. The company reports touchless claim rates above 90 percent and strong net collection rates, with customers like Talkiatry citing large reductions in manual work while volume climbed.

For a digital health company that wants billing to run in the background without a billing team, Candid's model fits. You push patient and encounter data through the API, claims run through the rules engine, and rejection data feeds back into your own workflows.

The platform has real depth on the engineering side. Its API reference covers encounter submission, eligibility, credentialing data, patient records, and organizational configuration. If your company already maintains integrations, Candid is a serious option, and its investors (Oak HC/FT, 8VC, First Round Capital, Y Combinator) reflect conviction in the API-first RCM category. Candid has also moved upmarket toward multi-site provider groups and MSOs alongside its original digital-health base.

Where a billing-company-first platform fits better

Independent billing companies do not look like Candid's core customer. They are not building a product around billing. They run a service business that manages five, ten, or thirty practices for clients who are mostly unaware of which clearinghouse processes their claims.

That shape needs different software:

• A shared work queue that surfaces denials and aging across all client practices in one view, not a set of per-client API dashboards.
• ERA review a biller can work through in a morning, rather than a pipeline that assumes automation handled it.
• Multi-practice claim submission and tracking where the billing company is the operator, not a practice admin.
• Pricing that does not change when a client practice adds providers, and that falls on a published graduated scale as your total book of practices grows.

Candid does list billing services as a segment on its billing services page and offers white-label capabilities. The product is still designed around API integration, so a billing company adopting Candid either builds and maintains the connection layer or relies on out-of-the-box EHR integrations that may not match a varied client mix. Medi is built for the billing company as the direct operator, with no integration to maintain, pricing that does not change when a client practice adds providers, and a published graduated rate that falls as the billing company's total practice count grows.

Is Medi always a better fit than Candid Health?

No. This is about fit, not quality.

If you are a digital health company that wants billing handled as infrastructure and you have engineers on staff, Candid is a well-funded, technically capable option with a track record at companies like Nourish, Talkiatry, and Tia.

If you are a high-volume operation with developers who can own an API integration, Candid's percentage model could come out cheaper than per-transaction pricing at scale.

Where the fit breaks down is the independent billing company that needs to log in every morning and work a shared queue across unrelated clients. The operator workflows, multi-practice navigation, and biller-centric screens those companies depend on are not the design center of an API-first platform. If you are choosing between Candid and Medi, the first question is what kind of organization you are: a digital health company with engineers, or a billing company with billers.

Other comparisons billing companies look at

• Medi vs Tebra — practice-first EHR plus billing, priced per provider.
• Medi vs AdvancedMD — per-provider PM platform with a central billing office offering.
• Billing company software evaluation guide — the side-by-side criteria that matter regardless of shortlist.

Frequently asked questions

Is Candid Health designed for independent billing companies?

Candid lists a billing services segment and describes itself as usable by third-party billing organizations. In practice the platform is API-first, built around direct integration with the systems that generate encounters. An independent billing company would need either a compatible EHR integration or development resources to connect its workflow. The core design targets digital health companies and provider groups that want billing embedded in their product, not billing companies that operate across many unrelated client practices.

How is Candid Health priced?

Candid does not publish pricing. Third-party reporting describes a percentage of collections, with a higher rate when Candid also handles coding, and implementation costs quoted separately. Actual pricing varies by volume, specialty, and contract, so you have to request a quote. The figures here come from Out-of-Pocket Health analysis, not from Candid's own published rates.

Does Candid Health replace human billers?

No. Candid's pitch is reducing manual work, not removing the billing function. Customers report large drops in manual effort, which means billers spend less time on rework and follow-up, not that the role disappears. Claims that need judgment, appeals that need a narrative, and payer-specific denial categories still involve people. Candid's engine catches errors before submission and automates resubmissions where rules apply cleanly. Medi takes the same boundary: it surfaces and organizes work for billers rather than acting without human review.

What engineering work does Candid require to go live?

At minimum you integrate with the encounter submission API to start sending claims. Which additional endpoints you build (patient records, eligibility, exports) depends on how much of your billing stack you want Candid to own. The integration guide covers the scope. For organizations without a developer, Candid offers pre-built integrations with certain EHR platforms, though coverage varies by system.

Who are Candid Health's typical customers?

Publicly documented customers include Talkiatry, Nourish, Tia, and Cortica: tech-enabled, venture-backed healthcare companies with engineering teams. Candid has also said it is moving upmarket to multi-site provider groups and MSOs. A billing company managing a portfolio of unrelated practices is not the profile the platform centers on.

How does Candid's AI compare to what a billing-company platform offers?

Candid's AI focuses on pre-submission validation, autocorrection, and first-pass acceptance, learning from aggregate claims data to build predictive rules. That is real depth for one organization running high volume through a single payer mix. A billing-company-first platform makes a different bet: the biller is the first-class operator who applies judgment across diverse payers, specialties, and client-specific rules, and the software surfaces work to act on rather than processing claims without review. Neither bet is wrong. They assume different operators.