Quick answer: Check seven things: a branded client login, tenant isolation, a provisioning API, client billing, per-client concurrency, phone-number ownership and exit terms. On the vendors’ own pages as at 26 September 2026, HighLevel documents 6 of the 7, Synthflow 5, Retell AI 4 and Vapi 3. Synthflow removed in-product rebilling on 15 September 2026.
If you have typed “can you white label Vapi” or “can you white label Retell AI” into a search box, the short answer is that four platforms agencies commonly compare (Vapi, Retell AI, Synthflow and HighLevel) take four different approaches. The White-Label Seven below shows which approach suits which kind of agency. Every vendor fact on this page comes from that vendor’s own pricing page, documentation or help centre, read on 26 September 2026 and linked in the source table at the end. Stripe, LeadConnector, Twilio and Australia’s carriage rules come in where they change the answer.
What should I check before choosing a white label AI voice agent platform?
“White label” means three different things. The first is a branded surface: your client logs into a dashboard on your domain with your logo. The second is tenancy: each client’s agents, numbers, recordings and usage are kept separate from every other client’s. The third is commercial plumbing: you can bill the client, mark the cost up, cap what they use and walk away cleanly when they leave. A platform can be excellent at one and silent on the other two.
The White-Label Seven splits those into checks you can confirm on a vendor’s own pages before you sign, each with the failure it prevents.
| # | Check | The question to ask the vendor | What a missing answer costs you |
|---|---|---|---|
| 1 | Branded client surface | Can my client log in on my domain, with my logo, and never see your name? | You build and host your own front end, or your client sees the vendor and can go direct |
| 2 | Tenant isolation | Is each client a separate workspace, with its own agents, numbers, recordings, keys and user roles? | One client’s login can expose another client’s calls; usage arrives as one number you have to divide by hand |
| 3 | Provisioning API | Can my onboarding flow create a client tenant without a person clicking through a dashboard? | Every new client adds manual setup time, which caps how fast you can grow |
| 4 | Client billing and markup | Can the platform charge my client, at my price, or do I invoice outside it? | You rebuild billing in your own Stripe or accounting system and reconcile usage every month |
| 5 | Per-client concurrency | Is the simultaneous-call limit set per client, or pooled across my whole account? | One client’s outbound campaign can take the lines another client’s inbound calls need |
| 6 | Phone-number ownership | If I buy a number and give it to a client, who owns it when they leave, and whose name is on the phone bill? | A dispute over a number the client’s customers already dial; possible carriage obligations in Australia |
| 7 | Exit terms | What can I export, how long do you keep raw call data, and what does deleting a tenant do? | Recordings and configuration lost at deletion, or kept longer than your client agreed |
Check 5 is the one agencies discover in production: a pooled concurrency limit looks generous until two clients run campaigns in the same hour. Check 6 is the one that becomes a dispute, because the client already thinks of the number as theirs. In Australia it also has a regulatory side: if your business name is on the bill for the phone service, our carriage service provider test for AI voice resellers explains why you should assume you are a CSP and which obligations follow.
Can you white label Vapi?
Yes, but through the API rather than through a rebranded Vapi dashboard. Vapi’s own FAQ answers the question directly: “Do you support white-labeling? Yes, Vapi is a API-first platform. This means can bake it in your product without using Vapi branding.” The same FAQ answers multi-tenancy the same way: “The integration model of multi-tenancy is similar to Stripe where your application logic will take care of it.” In other words, the branded surface (check 1) and the tenant model (checks 2 and 3) are yours to build.
Vapi does have a unit of separation, the organisation. Its billing docs describe creating organisations “to separate workspaces, teams, or environments under the same subscription”. The pricing page includes 2 organisations on Usage only and on Core ($29 a month), 10 on Pro ($999 a month minimum), and charges $20 per additional organisation per month. The detail that matters for an agency is on the organisations page: “concurrent call lines are shared by all organizations in the subscription.” The package sets the pool (4 concurrent calls on Usage only, 10 on Core, 30 on Pro, extra lines at $10 per line per month), and every client organisation draws from that same pool. That answers check 5: on Vapi, concurrency is pooled unless your own application enforces per-client limits.
For agencies that don’t want to build a front end, Vapi’s own documentation lists four third-party white-label dashboards that sit on a Vapi API key: ChatDash, Vapify, Voicerr AI and VoiceAIWrapper. Vapi’s page for Vapify, for example, describes “your logo, your domain, your own email address” and a rebilling feature: “Set custom monthly pricing and markup voice assistant prices”. Those are claims made by the third parties on pages Vapi hosts. Each dashboard is a separate contract to check with that company; the Vapi layer underneath, including shared concurrency, stays the same.
On exit (check 7), Vapi’s docs state that deleting an organisation “permanently removes its contents”, including “assistants, phone numbers, files, and other resources”. Raw data retention is 14 days on Usage only, 30 days on Core and 180 days on Pro. If a client agreement promises recordings for longer than that, you need to export them yourself. Vapi’s FAQ also states that it does not support on-premise deployments.
Can you white label Retell AI?
Retell documents strong tenant isolation but no rebranded client dashboard. Its workspace documentation says “Each workspace is a fully isolated boundary in Retell”. A workspace separates agents, conversation flows, knowledge bases, voices and phone numbers, and also API keys, members and roles, webhooks, telephony settings, and billing. Each workspace has its own “payment method, credit balance, invoices, and usage totals”. The same page recommends one workspace per client for an agency or reseller. That is an unusually explicit answer to check 2, and it makes check 4 simpler: usage for each client is already separated at the workspace level.
Retell answers check 3 in the negative itself: “Retell does not expose an API to programmatically create, delete, or switch workspaces, so per-tenant workspace provisioning has to be done by hand in the dashboard.” For a multi-tenant SaaS product, Retell’s docs suggest the opposite pattern: keep all resources in one workspace, “model your tenants in your own application database”, and attribute usage with call metadata. That is the same build-it-yourself approach Vapi describes.
Check 5 favours Retell. Its concurrency docs say limits “apply per workspace, not per account”, and that “traffic in one workspace does not consume slots in another”. Pay-As-You-Go workspaces get a quota of 20 concurrent calls. The pricing page puts it slightly differently (“Every account includes 20 concurrent calls for free”) and prices extra capacity at $8 per concurrent call per month. If you plan dozens of client workspaces, ask Retell which wording applies to your bill.
On check 1: no rebranded client dashboard or custom-domain setting is documented on Retell’s workspace docs, docs index, Solution Partner Program page, pricing page or public changelog as at 26 September 2026. We searched each of those pages and confirmed each had loaded properly by finding text we knew was on it. Retell’s support portal at support.retellai.com is a JavaScript application that returned no readable content to our fetch, so that source could not be checked. Retell’s Solution Partner Directory does list partners that describe white-label platforms built on Retell’s API, on the agency’s own domain and branding. Like the dashboards on Vapi’s integration pages, those are third-party claims on a page Retell hosts, and each is a separate contract to check. Retell’s own agency pricing guide, published 25 September 2026, treats white-labelling as a positioning choice for the agency (“If the client never sees the underlying platform…”), not as a product setting. On check 7, deleting a workspace permanently deletes all its data and triggers a final invoice.
Synthflow’s agency white-label: what changed on 15 September 2026
Synthflow fully documents five of the seven checks and part of a sixth, and it made one important change eleven days before this page was written. Its agency page promises “Own White Label Brand and Domains” and “Multi-Tenant Architecture”. Its docs back that up. There is a custom-domain flow (DNS record, allow up to 24 hours to propagate), branding settings, sub-accounts that give “each client their own workspace under your agency account”, per-sub-account permissions, and a Subaccounts API that covers create, update, delete, get and list.
Check 5 is partly documented. Synthflow’s Limits & Capabilities page sets included minutes, concurrent calls, maximum agents, allowed agent types and workflow runs per sub-account, and you can set defaults for new sub-accounts and override them for existing ones. Those are caps, not reserved capacity: Synthflow’s concurrency page says “Concurrency is set by your plan, and your accounts and subaccounts share the main account’s total.” A per-client cap stops one client taking every line, but it does not keep a quiet client’s lines free. Check 6 is written down in plain terms: a number the parent account bought and transferred to a sub-account “is still yours, and the subaccount cannot delete it”, and it goes back to the main workspace if the sub-account is deleted. For check 7, deleted sub-accounts “cannot be restored”, manual CSV export works per agent, and automatic daily exports are Enterprise-only.
The change affects check 4. Synthflow’s migration guide says in-product reselling “was removed on September 15, 2026”. That covers the custom pricing plans, the Plan & Billing and Usage pages for sub-accounts, and the Stripe integration. Anyone billing clients through Synthflow now needs their own Stripe or other billing system. Sub-accounts, permissions, branding, custom domains and products “continue to work”, and Limits & Capabilities is “now the primary way to cap subaccount usage”. One behaviour to watch: with Auto-overage switched on, a sub-account keeps calling past its allocation and “the extra minutes are billed to your account at your per-minute rate”.
Access also changed. Synthflow’s concurrency page says the earlier Pro, Growth and Agency plans “are no longer available for new subscriptions”. Its pricing page lists a single Enterprise option: “Enterprise contracts start at $30,000 annually.”
HighLevel: white-label comes from the CRM layer
HighLevel is a white-label CRM and marketing platform that includes its own Voice AI, not a voice platform with white-label added on. That changes which checks it passes. The pricing page covers check 1 (“you have the ability to white-label the desktop web app, meaning your clients see your logo and branding when they log in”, plus custom domains) and check 2 (3 sub-accounts on Starter at $97 a month, unlimited on Unlimited at $297 and Agency Pro at $497). Agency Pro adds “SaaS Mode”, “Automated Sub-Account Creation” and “Rebill Phone & Email with Markup”, which covers checks 3 and 4.
For voice specifically, HighLevel’s AI pricing article (modified 25 September 2026) states that agencies “must be on the $497/month plan to rebill AI Employee usage”. It gives its own Pay-Per-Use worked example: a 10-minute Voice AI call using OpenAI TTS costs $0.60 (Voice Engine $0.045 a minute plus TTS $0.015 a minute) “before LLM token usage and Phone System charges”. The wallet model means HighLevel always charges the agency wallet first, and your own Stripe account then charges the sub-account if rebilling is on.
Check 5 is where HighLevel is candid. Its inbound rate-limit article says HighLevel “does not publish a separate per-number, per-location, or per-agent concurrent-session ceiling” for phone-number Voice AI, and warns that this “should not be interpreted as a guarantee of unlimited simultaneous sessions.” What it does publish is an inbound threshold of more than 200 calls per minute to or from a Voice AI number, and 20 concurrent browser-based calls per location for the Voice AI chat widget. Checks 6 and 7 are covered by the sub-account transfer article. When a sub-account is ejected to a new agency, phone numbers move automatically only if both agencies are on LeadConnector phone. Where either side uses Twilio, “No phone numbers will be transferred over by HighLevel.” SaaS mode is switched off before the transfer, and domains have to be re-added.
White-label AI voice platforms compared: the White-Label Seven tally
Each cell records what that vendor’s own pages document, read on 26 September 2026. “Documented” means the vendor describes a working mechanism. “Vendor says no” means the vendor states the capability is absent. “Not documented” is used only after the route check described above.
| Check | Vapi | Retell AI | Synthflow | HighLevel |
|---|---|---|---|---|
| 1. Branded client surface | Via your own product on the API (FAQ); or one of 4 third-party dashboards listed in Vapi’s docs | Not documented as a Retell setting on the pages listed above, as at 26 September 2026; third-party white-label platforms listed in Retell’s partner directory | Documented: custom domain and branding | Documented: white-labelled web app, custom domains; white label mobile app add-on $497/month |
| 2. Tenant isolation | Organisations: 2 included (Usage only, Core), 10 (Pro), +$20/org/month; FAQ puts tenancy in your application | Documented: workspace is “a fully isolated boundary” | Documented: sub-account per client, with permissions | Documented: sub-accounts, 3 on Starter, unlimited on $297 and $497 plans |
| 3. Provisioning API | Tenants live in your own application (FAQ) | Vendor says no: no API to create, delete or switch workspaces | Documented: Subaccounts API | Documented: automated sub-account creation on Agency Pro ($497) |
| 4. Client billing and markup | Not in Vapi itself; third-party dashboards describe rebilling | Per-workspace payment method and invoices; no markup feature documented on the pages listed | Removed 15 September 2026; bill outside Synthflow | Documented: rebill with markup on $497; AI Employee usage rebilling needs $497 |
| 5. Per-client concurrency | Pooled: lines “shared by all organizations”; 4 / 10 / 30 by package, $10/line add-on | Documented: per workspace; 20 on Pay-As-You-Go; $8/concurrent call/month extra | Partial: concurrent-call cap per sub-account, inside a total that “your accounts and subaccounts share” | Phone Voice AI: no ceiling published (vendor’s own statement); widget 20 per location |
| 6. Number ownership on exit | Deleting an organisation removes its phone numbers | Numbers belong to the workspace; Retell’s guide says settle release terms at signing | Documented: parent keeps transferred numbers | Documented: numbers move only if both agencies are on LeadConnector |
| 7. Exit terms | Raw data retention 14 / 30 / 180 days by package; deletion permanent | Deletion permanent, final invoice raised | Deletion permanent; CSV export; daily exports Enterprise-only | Documented: eject-to-new-agency flow, with listed exclusions |
| Documented of 7 | 3 (2, 6, 7); 1, 3 and 4 are left to you; 5 is pooled | 4 (2, 5, 6, 7); 1 not documented; 3 vendor says no; 4 partial | 5 (1, 2, 3, 6, 7); 5 partial (per-sub-account caps on a shared total); 4 removed | 6 (1, 2, 3, 4, 6, 7); 5 not published |
The tally measures what each vendor documents, not product quality. A low number for an API-first platform like Vapi means you own more of the build, not that the platform is weaker.
What 25 client organisations cost on Vapi before a single minute
Agencies model per-minute costs and forget per-tenant costs. Here is the arithmetic for 25 clients, each in their own Vapi organisation, from Vapi’s published pricing:
- Core package: $29 a month, 2 organisations included. The other 23 organisations cost 23 × $20 = $460. Total: $29 + $460 = $489 a month, with 10 concurrent lines shared by all 25 clients, or 0.4 lines per client on average.
- Giving each client an average of 2 lines needs 50 lines: 40 extra × $10 = $400, which brings Core to $889 a month. The lines are still one shared pool, so a busy client can use more than 2.
- Pro package: at least $999 a month, 10 organisations and 30 lines included. 15 extra organisations × $20 = $300, and 20 extra lines × $10 = $200. Total: at least $1,499 a month.
- The single-organisation pattern that Vapi’s FAQ describes avoids the $460 or $300 in organisation fees. The trade-off is that tenant isolation then depends on your own code.
For comparison, HighLevel’s Agency Pro plan is $497 a month with unlimited sub-accounts, and its AI Pay-Per-Use option has no monthly fee, so 25 clients cost $497 a month before usage. Agency Pro is the plan HighLevel requires to rebill AI Employee usage; its $297 Unlimited plan also includes unlimited sub-accounts, without SaaS Mode or markup on rebilled phone and email. Adding AI Employee Unlimited to all 25 sub-accounts (25 × $97 = $2,425) brings that to $2,922 a month, and that plan includes unlimited Voice AI (inbound, outbound and widget), subject to HighLevel’s fair-use policy, with Phone System charges billed separately. On Synthflow, the published entry point is an Enterprise contract from $30,000 a year, which averages $2,500 a month; Synthflow says final pricing is “scoped around call volume, concurrency, telephony setup, integrations, security needs, and launch support”, so what that figure covers depends on the contract. The Vapi totals and HighLevel’s $497 include no minutes or telephony; HighLevel’s AI Employee Unlimited includes Voice AI usage, subject to fair use, but not Phone System charges; and none of these figures includes the time spent building what the platform leaves to you. To model the per-minute side, our breakdown of voice AI cost per minute, realtime versus pipeline shows the inputs and the method.
Which white label route fits your agency: the threshold table
| If this describes you | Route that fits | Why, from the vendor pages |
|---|---|---|
| You have developers and want your own product, not a reskinned dashboard | API-first: Vapi or Retell AI | Both document building tenancy in your own application database |
| You need a branded client login this month and have no front-end team | Self-serve white-label: HighLevel Agency Pro, or Synthflow on an Enterprise contract | Both document custom domain, branding and sub-accounts |
| Each client needs guaranteed simultaneous-call capacity | Per-tenant concurrency: Retell workspaces | Retell gives each workspace its own quota; Synthflow caps each sub-account but its sub-accounts share the account total; Vapi pools lines across organisations; HighLevel publishes no phone Voice AI ceiling |
| You onboard clients through your own signup flow | Provisioning API: Synthflow Subaccounts API or HighLevel SaaS Mode | Retell states workspace creation is manual |
| You want the platform to charge clients at your price | HighLevel on $497 | Synthflow removed in-product rebilling on 15 September 2026 |
| A regulated client needs the model on their own infrastructure | A private deployment option (see the Zian section below) | Vapi’s FAQ says it does not support on-premise deployments |
Who each option is wrong for
Vapi is wrong for an agency without developers. Vapi’s own answer to white-labelling is to build it into your own product. The third-party dashboards close that gap, but you then depend on two vendors and have two contracts to read. Vapi is also a poor fit if you promise clients their own dedicated capacity, because lines are shared across organisations.
Retell AI is wrong for an agency that onboards many small clients quickly. Workspaces are well isolated and concurrency is set per workspace. But every client workspace is created by hand, and no branded client login is documented in Retell itself. It suits a few larger clients, each with a scoped workspace.
Synthflow is wrong for a small agency that wanted the old self-serve agency plan, and wrong for anyone who relied on it to bill clients. It documents tenant isolation, a provisioning API and per-sub-account concurrency caps (checks 2, 3 and part of 5), but the published route is now an Enterprise contract, and billing happens in your own systems.
HighLevel is wrong for an agency that needs guaranteed voice concurrency, or already has its own CRM. Its white-label strength comes from the CRM layer, and HighLevel’s own help centre says it does not publish a phone Voice AI concurrency ceiling. If your clients already run on HubSpot or Salesforce, you would be adding a second CRM to get a white-label voice layer.
Where Zian fits: a white label partnership, not a settings page
Zian publishes its partner model on its solutions page as three tracks: Affiliate (refer opportunities for a commission), White Label (“Offer Zian AI under your brand with dedicated onboarding and support”, described as “Your brand, our stack”, with co-selling and enablement) and Joint Venture (bring an audience or a niche workflow; Zian brings engineering, infrastructure and go-to-market). Zian is in partnership-application beta, so these are agreed through an application, not switched on in a self-serve dashboard. The seven settings above are not published as product toggles on zian.ai, so a partner should ask Zian the same seven questions this page asks everyone else, and get the answers in writing.
Two things Zian does publish matter for the threshold table above. The first is private model deployment on customer infrastructure. For a regulated client whose data cannot sit in a shared cloud, private AI deployment for AI sales agents sets out what running agents on your own infrastructure involves. The second is API and CRM integrations with HubSpot, Salesforce, HighLevel and Zapier; the native, Zapier or API integration patterns page compares them. The operating context is multi-tenant by nature. Zian has run outbound acquisition since 2017 and at one point operated campaigns for roughly a hundred businesses in the same vertical simultaneously, which is exactly the situation where checks 5 and 6 stop being theoretical.
Frequently asked questions
Can you white label Vapi?
Yes, through its API. Vapi’s own FAQ says you can build it into your product without Vapi branding, and that multi-tenancy is handled by your application logic, similar to Stripe. Vapi’s docs also list four third-party white-label dashboards: ChatDash, Vapify, Voicerr AI and VoiceAIWrapper. Concurrent call lines are shared across all organisations in a subscription.
Can you white label Retell AI?
Retell documents one isolated workspace per client, with separate keys, roles, numbers and invoices, and concurrency limits set per workspace. It states that there is no API to create workspaces, so each one is set up by hand. No rebranded client dashboard or custom-domain setting in Retell itself is documented on the Retell pages we checked as at 26 September 2026; Retell’s partner directory lists third parties that build white-label platforms on its API.
Does Synthflow still offer white label for agencies?
Yes for branding, custom domains, sub-accounts, permissions and per-sub-account limits. No for billing: Synthflow’s reselling migration guide says in-product reselling through Stripe was removed on 15 September 2026, so you now bill clients yourself. Its pricing page lists Enterprise contracts from $30,000 a year.
Who owns the phone number if my client leaves?
It depends on the vendor and on who bought the number. On Synthflow, a number the parent account bought stays with the parent. On HighLevel, numbers move with an ejected sub-account only if both agencies use LeadConnector phone. On Vapi, deleting an organisation removes its numbers. Put the release terms in the client contract at signing.
Do I need to be a carriage service provider to resell AI voice agents in Australia?
Not automatically. The question usually turns on whose name is on the bill for the phone service. Our carriage service provider test for AI voice resellers works through the threshold, the obligations that follow and who to ask for a ruling.
Does Zian offer a white label programme?
Yes. Zian’s solutions page lists a White Label partnership (“Offer Zian AI under your brand with dedicated onboarding and support”) alongside Affiliate and Joint Venture tracks. Zian is in partnership-application beta, so terms are agreed through the application, not set in a self-serve dashboard.
Where every figure on this page comes from
| Figure | Who published it | Link | Date read |
|---|---|---|---|
| Core $29/month; Pro $999/month minimum; 2 / 10 organisations; 4 / 10 / 30 concurrent calls; $20 per extra organisation; $10 per extra line; 14 / 30 / 180 days raw data retention | Vapi | vapi.ai/pricing | 26 September 2026 |
| Concurrent call lines shared by all organisations in a subscription; organisation deletion removes phone numbers | Vapi | docs.vapi.ai, Manage organizations and capacity | 26 September 2026 |
| White-labelling via API; multi-tenancy in your application; no on-premise | Vapi | docs.vapi.ai/faq | 26 September 2026 |
| 4 third-party white-label dashboards (ChatDash, Vapify, Voicerr AI, VoiceAIWrapper) | Vapi (integration pages) | docs.vapi.ai/providers/vapify and sibling provider pages | 26 September 2026 |
| Workspace isolation; no API to create, delete or switch workspaces | Retell AI | docs.retellai.com/accounts/workspace | 26 September 2026 |
| 20 concurrent calls per Pay-As-You-Go workspace; limits per workspace | Retell AI | docs.retellai.com/deploy/concurrency | 26 September 2026 |
| 20 free concurrent calls per account; $8 per extra concurrent call per month | Retell AI | retellai.com/pricing | 26 September 2026 |
| Settle number release terms at signing (agency setup guide, published 25 September 2026) | Retell AI | retellai.com/blog/running-voice-ai-for-clients | 26 September 2026 |
| In-product reselling removed 15 September 2026 | Synthflow | docs.synthflow.ai/reselling-migration-guide | 26 September 2026 |
| Enterprise contracts from $30,000 annually; final pricing scoped around call volume, concurrency, telephony setup, integrations, security needs and launch support | Synthflow | synthflow.ai/pricing | 26 September 2026 |
| Earlier Pro, Growth and Agency plans closed to new subscriptions | Synthflow | docs.synthflow.ai, Concurrency | 26 September 2026 |
| Per-sub-account minutes, concurrency and agent limits | Synthflow | docs.synthflow.ai/limits-and-capabilities | 26 September 2026 |
| DNS propagation up to 24 hours for custom domain | Synthflow | docs.synthflow.ai/custom-domain | 26 September 2026 |
| Transferred numbers stay with the parent account | Synthflow | docs.synthflow.ai/phone-numbers | 26 September 2026 |
| $97 / $297 / $497 plans; 3 sub-accounts on Starter, unlimited on $297 and $497; SaaS Mode and rebilling with markup on $497; $97 per sub-account AI Employee Unlimited; $497 white label mobile app | HighLevel | gohighlevel.com/pricing | 26 September 2026 |
| $0.045/min Voice Engine; $0.015/min OpenAI TTS; $0.60 per 10-minute Pay-Per-Use example; $497 plan to rebill AI Employee usage; AI Employee Unlimited includes unlimited Voice AI, subject to fair use; Pay-Per-Use has no monthly fee | HighLevel | help.gohighlevel.com, AI product pricing (modified 25 September 2026) | 26 September 2026 |
| More than 200 calls/minute Voice AI inbound threshold; 20 concurrent widget calls per location; no phone Voice AI concurrency ceiling published | HighLevel | help.gohighlevel.com, inbound call rate limits | 26 September 2026 |
| Numbers transfer on eject only when both agencies are on LeadConnector | HighLevel | help.gohighlevel.com, eject sub-account | 26 September 2026 |
| Agency wallet always charged for sub-account usage; Agency Stripe charges the sub-account when rebilling is on | HighLevel | help.gohighlevel.com, rebilling, reselling and wallets | 26 September 2026 |
| Accounts and subaccounts share the main account’s concurrency total | Synthflow | docs.synthflow.ai, Concurrency | 26 September 2026 |
| Subaccounts API (create, list, read, update, delete); deleted sub-accounts cannot be restored; parent-owned numbers return on deletion | Synthflow | docs.synthflow.ai/subaccounts | 26 September 2026 |
| Manual CSV export per agent; automatic daily exports Enterprise-only | Synthflow | docs.synthflow.ai/export-call-data | 26 September 2026 |
| “Own White Label Brand and Domains”; “Multi-Tenant Architecture” | Synthflow | synthflow.ai/industry/agency | 26 September 2026 |
| Third-party white-label platforms built on Retell’s API | Retell AI (partner listings) | retellai.com/partners | 26 September 2026 |
| “If the client never sees the underlying platform…” (agency pricing guide, 25 September 2026) | Retell AI | retellai.com/blog/how-agencies-price-ai-voice-agents | 26 September 2026 |
| Worked totals ($489, $889, $1,499, $497, $2,922, $2,500 a month) | Zian, calculated from the vendor figures above | This page | 26 September 2026 |
| White-Label Seven tally (3 / 4 / 5 / 6 of 7) | Zian, from the vendor pages above | This page | 26 September 2026 |
| Three partner tracks: Affiliate, White Label, Joint Venture | Zian | zian.ai/solutions/ (Partner Programme section) | 26 September 2026 |
| Outbound since 2017; roughly a hundred businesses in one vertical at once | Zian (first-party, owner-released) | zian.ai | 26 September 2026 |