Best Synthflow Alternatives in 2026: 8 Voice AI Platforms With Real Prices
Synthflow sells only annual contracts starting at $30,000 a year, or $2,500 a month, so most teams leave over arithmetic. Retell AI, Vapi, Bland and ElevenLabs Agents all cost between about $812 and $1,400 a month at 10,000 minutes on their published rates.
Key takeaways
- Synthflow now sells only annual contracts starting at $30,000 a year, or $2,500 a month, after closing its self-serve Pro, Growth and Agency plans to new subscribers.
- Retell AI is the closest self-serve Synthflow alternative, charging $0.07 to $0.31 a minute with $10 of free credit and 20 concurrent calls included, about $1,144 at 10,000 minutes.
- Vapi charges a $0.05-a-minute platform fee plus provider costs at cost, which its calculator puts at $819 to $1,290 for 10,000 minutes a month.
- ElevenLabs Agents bills $0.08 a minute on every plan, from a free tier with 15 minutes to Business at $990 for 12,375 minutes, with telephony billed separately.
- PolyAI's AWS Marketplace listing sets a 12-month minimum of 500,000 minutes for $175,000, or $14,583 a month, which costs more to enter than Synthflow's floor.
The best Synthflow alternatives are Retell AI, Vapi, ElevenLabs Agents and PolyAI, because Synthflow now starts at $30,000 a year. Retell AI is the closest self-serve match for teams who liked Synthflow's builder, Vapi suits developers who want to choose every component, ElevenLabs Agents suits teams who care most about how the agent sounds, at a flat $0.08 a minute, and PolyAI is the option if what you actually want is a vendor that runs the agent for you. For the product itself, see our Synthflow complete guide; this page is about leaving it.
Which Synthflow alternative fits your situation?
| Your situation | Look at | Why |
|---|---|---|
| You liked the no-code builder but not the $30,000 floor | Retell AI | Pay as you go, 20 concurrent calls included, $10 free credit |
| You have developers and want to pick the STT, LLM and voice yourself | Vapi | $0.05 a minute platform fee, providers at cost |
| You run high-volume outbound campaigns | Bland | Plans built around calls per day, from $0 platform fee |
| Voice quality matters most and you want one flat minute rate | ElevenLabs Agents | $0.08 a minute on every plan, speech and voice included; model and telephony extra |
| You're a sales team converting inbound leads by phone, SMS and email | Thoughtly | Flat plan from $500 a month with unlimited minutes |
| You also need a business phone system for human agents | CloudTalk | AI agent minutes sit next to per-seat phone plans |
| You're a small business answering your own phone | Goodcall | Priced per unique caller, not per minute |
| You want a vendor to build, run and tune the agent | PolyAI | Managed service; its AWS listing starts at $175,000 a year |
Seven of the eight publish a price you can check today. The exception, PolyAI, publishes one on AWS Marketplace, which we cover below.
Why do teams leave Synthflow?
We read every low-star review on Synthflow's Trustpilot page on September 19, 2026. Synthflow's G2 reviews (4.5 from 1,015, checked September 19, 2026) don't add complaint detail beyond the rating, and the Reddit thread that ranks for this search wouldn't load, so the complaints below come from Trustpilot and from Synthflow's own pricing page. Four themes recur.
1. The price moved out of reach for small teams. Synthflow's pricing page now lists one Enterprise tier: "Enterprise contracts start at $30,000 annually." Its billing docs say "Synthflow uses sales-led pricing," and the older Pro, Growth and Agency plans are closed to new subscriptions. Users felt the squeeze before the floor arrived. Dan Miller (US, August 18, 2025) wrote on Trustpilot that "the tools you actually need are locked behind their absurdly overpriced enterprise plan."
2. Agents and actions misbehaving in production. A reviewer listed as Small Business Owner (US, May 4, 2026) reported that "the AI agents suffer from severe hallucinations" and described "frequent failures with basic integrations like SMS triggers and webhooks." For a voice agent, a failed webhook usually means a booking or CRM update that never happened.
3. Reaching a person for support or billing. Tyler Combs (US, April 11, 2026) wrote that he "cancelled my subscription over 5 moths ago and continue to get billed each month" and that "you cant get a human on the phone." Another reviewer (Canada, December 24, 2025) left two words: "No support." The same Small Business Owner review said support "offers generic troubleshooting steps."
4. White-label gaps for agencies. A reviewer named Leo (US, April 26, 2025) bought the white-label service and found that "Synthflow's logos and branding remained visible throughout the platform." Synthflow replied a month later asking for account details. Agencies reselling voice agents to their own clients hit this one hardest. We didn't verify white-label features for any of the eight alternatives below, so if this is your reason to leave, ask each vendor for a branded demo workspace before you sign.
Read these alongside the ratings. Synthflow holds 4.1 on Trustpilot from 215 reviews, but only 7 were posted in the last 12 months, so the score mostly reflects the self-serve era. Its pricing page displays "4.5, 1,000+ reviews on G2," which G2 confirms: 4.5 from 1,015 reviews, checked September 19, 2026. On Gartner Peer Insights it has 4.3 from 4 ratings. Capterra still lists a "Synthflow AI Receptionist" at "$29/Per Month" with no reviews, a leftover from the old plans.
Who should stay on Synthflow?
Plenty of teams should. If you already have a legacy Pro, Growth or Agency plan, Synthflow's concurrency docs say those plans "remain supported for existing customers," so nothing forces a move. Volume is the other reason. The break-even we use throughout is $30,000 ÷ 12 = $2,500 a month, then $2,500 ÷ $0.08 ≈ 31,000 minutes, where $0.08 is the last per-minute rate Synthflow published itself, in an August 2025 blog post. That is our assumption about what the contract buys; Synthflow doesn't publish its contract rates, and yours may differ. Below about 15,600 minutes a month, more than half of the floor pays for minutes you don't use. Above roughly 31,000, the floor stops being dead weight and becomes a normal contract, with an account team, launch support and SLA terms included. And Synthflow's metering is generous in ways the per-minute rivals aren't: its usage docs pool seconds across calls rather than rounding each call up, stop billing after a successful transfer, charge unanswered and voicemail-hangup calls a flat 5 seconds, and don't bill failed or user-canceled calls at all.
The homepage now sells to enterprises: the headline reads "Enterprise-Ready Voice AI Agents for Automated Phone Calls," and the only button in the hero is Get a demo. Stay, too, if your agent depends on Synthflow's Freshworks app or its GoHighLevel subaccount sync and you'd have to rebuild that plumbing. The switching section below lists what you'd have to rebuild.
How do the Synthflow alternatives compare on price and ratings?
Prices checked September 19, 2026 on each vendor's pricing page and rechecked for Synthflow, Retell, Bland and ElevenLabs on September 24, 2026; the Retell, Vapi, Bland and ElevenLabs figures also match our pricing posts on those vendors from the same day. Ratings are Trustpilot unless noted.
| Platform | Billing unit | Published entry price | Concurrency included | Rating (count) | Solves which complaint |
|---|---|---|---|---|---|
| Synthflow | Annual contract | From $30,000 a year | Scoped in contract | 4.1 (215) | n/a |
| Retell AI | Per minute, by component | $0.07–$0.31 a minute, $10 free credit | 20 calls, then $8 a line | 4.9 (814); G2 4.8 (2,638) | Price floor |
| Vapi | $0.05 a minute + providers at cost | Usage only, $5 credit; Core $29 a month | 4 (usage only), 10 on Core | 2.4 (15); G2 3.0 (3) | Control over a misbehaving agent |
| Bland | Plan fee + per minute | Start $0 + $0.14 a minute | 10 calls, 100 calls a day | 2.9 (2) | Outbound volume |
| ElevenLabs Agents | Plan minutes, $0.08 each | Free (15 min); Starter $6 | 4 (Free) to 40 (Business) | 3.0 (1,150, whole company) | Price floor |
| Thoughtly | Flat monthly plan | Flex from $500 a month | 10 calls on Flex | 4.0 (7) | Unpredictable bills |
| CloudTalk | Minute bundles + phone seats | AI Receptionist from $99 a month (200 min) | Not published | 4.1 (823, whole company) | Human backup and support |
| Goodcall | Per agent, per unique caller | Starter $79 a month (100 callers) | Not published | 2.5 (5) | Price floor for small businesses |
| PolyAI | Per minute, contract | AWS: $175,000 a year for 500,000 min | Scoped in contract | 3.5 (2) | Support (managed service) |
What does each alternative cost at 1,000 and 10,000 minutes?
We modeled an inbound line with a 4-minute average call, so 1,000 minutes is about 250 calls a month and 10,000 minutes about 2,500. Each row uses the cheapest combination the vendor sells that covers the volume, with the arithmetic shown.
| Platform | 1,000 minutes a month | 10,000 minutes a month | How we got there |
|---|---|---|---|
| Synthflow (estimate) | $2,500 | $2,500 | $30,000 floor ÷ 12; the contract rate isn't public |
| Retell AI | $116 | $1,144 | $0.114 a minute (GPT 5.4 mini, platform voice, Retell US telephony, knowledge base) + $2 a number; 2 numbers at 10,000 |
| Vapi | $82–$129 | $819–$1,290 | Vapi's own calculator defaults (Deepgram, OpenAI, ElevenLabs, Vapi telephony $0) |
| Bland | $140 | $1,400 | Start plan: minutes × $0.14, no platform fee; Build ($299 + $0.12) costs $1,499 at 10,000 |
| ElevenLabs Agents | about $81 + telephony | about $812 + telephony | 1,000: Creator $22 + 725 × $0.08; 10,000: Pro $99 + 8,762 × $0.08 + about $12 of Gemini 2.5 Flash |
| Thoughtly | from $500 | from $500 | Flex "starts at $500 per month," unlimited voice minutes |
| CloudTalk | from $349 | Quote | AI Specialist bundle (1,000 min); no USD bundle listed near 10,000 min |
| Goodcall (estimate) | $129 | $1,249 | Growth covers 250 callers; Scale $249 + 2,000 extra callers × $0.50, assuming every call is a new caller |
| PolyAI | $14,583 minimum | $14,583 minimum | AWS contract: $175,000 ÷ 12 for up to 500,000 minutes a year |
At 10,000 minutes the five self-serve platforms (Retell, Vapi, Bland, ElevenLabs and Goodcall) land between about $812 (ElevenLabs, before telephony) and $1,400 a month. Thoughtly's sales-led Flex plan starts lower, at $500, if its 10-call concurrency fits, and Synthflow's floor sits at $2,500.
Two rows need a caveat. ElevenLabs bills telephony separately: receiving calls on a local number costs $0.0085 a minute on Twilio's US price list, so roughly $86 at 10,000 minutes, which puts ElevenLabs at about $898 all in, above Vapi's $819 low end (Vapi's figure already includes its own telephony at $0). And Goodcall's row swings a lot with repeat callers; if half your calls come from people who called before, Scale drops to about $624.
The 8 best Synthflow alternatives
Retell AI
Retell is the nearest like-for-like swap: a hosted voice-agent platform with a visual builder and phone numbers, sold pay as you go. Its pricing page lists $0.07 to $0.31 a minute for voice agents and "$0.002+" per message for chat agents, with $10 in free credit, 20 concurrent calls included and $8 a month for each extra line. The $0.07 floor covers only voice infrastructure and the voice itself; a realistic support stack with GPT 5.4 mini, US telephony and a knowledge base comes to about $0.114 a minute.
The review signal is the strongest in this list: 4.9 on Trustpilot from 814 reviews and, when our sister post read G2 the same day, 4.8 from 2,638. Retell's Trustpilot profile is claimed and invites reviews, which lifts the count. The top G2 complaint tags were missing features, price and learning curve.
Retell's docs show the builder a Synthflow user would move to: the prompt on the left, functions, knowledge base and webhook settings in the middle, a test panel on the right, and a running cost estimate above the prompt ($0.115 a minute for the example agent, close to our $0.114 stack).
It answers Synthflow's price complaint directly: at 10,000 minutes you pay about $1,144 with no contract. The watch-out is that every component is its own line, and Retell's docs describe a surcharge on long prompts, so the bill depends on choices you make in the builder. Choose it if you want Synthflow's workflow without the sales call. Our Retell AI guide covers the product in depth.
Vapi
Vapi is the developer's option. You pay Vapi $0.05 a minute for hosting and pay the transcriber, model and voice providers at cost, either through Vapi or with your own keys. Its pricing calculator puts 1,000 minutes at $82 to $129 with its defaults, and 10,000 at $819 to $1,290. The usage-only tier gives $5 of credit, 4 concurrent calls and one free US inbound number; Core at $29 a month raises that to 10 lines, and HIPAA is a $2,000-a-month add-on.
Vapi is also sold on AWS Marketplace (listing prodview-tskr3mpoou7is, $0.05 a minute), and Vendr reports a median contract of $105,000 a year for larger buyers. Reviews are thin and mixed: 2.4 on Trustpilot from 15 reviews and 3 out of 5 from 3 reviews on G2.
Vapi fits the Small Business Owner's complaint better than any no-code tool, because when an agent hallucinates or a webhook fails you can see and change every layer: the prompt, the model, the tool call and the server that receives it. The cost of that control is engineering time. Choose it if you have developers and want the stack to be yours. See our Vapi guide.
Bland
Bland has priced by plan since December 2025. Its pricing page shows Start at $0.14 a minute with no platform fee, 10 concurrent calls and 100 calls a day, and Build at $0.12 a minute plus $299 a month, with 50 concurrent calls and 2,000 calls a day. Bland's billing docs add a Scale plan at $499 plus $0.11 a minute, which sits in the docs but not on the pricing page. Transfers cost $0.03 to $0.05 a minute, and each outbound attempt or failed call carries a $0.015 minimum.
Bland's plans are built around calls per day, which is the right unit for an outbound campaign and the wrong one for a support line with a Monday spike. Its Enterprise tier lists on-prem or VPC deployment, a BAA, SSO and data residency. Trustpilot has only 2 reviews (2.9), and G2 shows 11 reviews (5 out of 5, checked September 19, 2026).
The watch-out is telephony. Bland's pricing FAQ says telephony is billed separately at pass-through, while its docs say SIP cost is included on Bland's endpoints; ask which applies to your numbers. Choose it if you dial out at volume or need the agent inside your own cloud. Our Bland AI guide has the rest.
ElevenLabs Agents
ElevenLabs sells its agent platform on monthly plans where every included or extra minute works out to $0.08: Free (15 minutes), Starter $6 (75), Creator $22 (275), Pro $99 (1,238), Scale $299 (3,738) and Business $990 (12,375), per its Agents pricing page. The language model is billed at cost on top, from fractions of a cent a minute, and telephony comes from your own carrier.
At 10,000 minutes that comes to about $812 on Pro plus overage, or roughly $898 once telephony is added, so it isn't the cheapest option here: Vapi's calculator starts at $819 with telephony included. The reason to pick it is the voice. Voices are the company's core product, which matters if callers complain your agent sounds synthetic. Trustpilot's 3.0 from 1,150 reviews covers the whole company, mostly its text-to-speech tools.
The watch-out is how overage works for new self-serve customers: extra minutes come from a prepaid balance, and ElevenLabs' Pay As You Go docs say that if the balance hits $0 and Auto Top Up is off or fails, "your service will pause immediately." Set auto top-up before a support line depends on it. Choose it if voice quality leads your list and you want one flat rate per minute. See our ElevenLabs Agents guide.
Thoughtly
Thoughtly sells to revenue teams rather than support desks: its pricing page says it is "trusted by 1,000+ revenue teams reaching every inbound lead." The Flex plan "starts at $500 per month" and includes unlimited voice minutes, SMS and email, 10 concurrent calls, 200+ integrations and two-way CRM sync. Scale adds 100+ concurrent calls, SOC 2 Type II reports and a 99.9% uptime SLA; Enterprise adds HIPAA with a BAA.
A flat plan answers the complaint about bills you can't predict. It also moves the constraint from minutes to concurrency: 10 simultaneous calls on Flex is the real ceiling. Thoughtly's page also says "usage above your plan's monthly allowances is billed in arrears," which sits oddly next to "unlimited," so ask what the allowance is. Every plan starts with a sales conversation, and Trustpilot shows 4.0 from 7 reviews.
Choose it if you're converting inbound leads across phone, text and email and Synthflow was your lead-qualification agent. We don't have a Thoughtly guide yet.
CloudTalk
CloudTalk is a business phone system that added AI voice agents, which makes it a different kind of alternative: the AI agent and your human agents share one phone platform. In the USD plan table on its pricing page, the AI Receptionist starts at $99 a month for 200 minutes and the AI Specialist at $349 for 1,000 minutes; phone plans run $25, $29 and $49 a user a month billed annually.
The Voice Agents tab rendered for us in euros, with different bundles: a free Receptionist with 50 minutes, the Specialist at €349 for 1,000 minutes, and custom pricing above 10,000 minutes "as low as €0.15 per minute." Check which bundle your region is offered. At $349 for 1,000 minutes, the Specialist works out to about $0.35 a minute, roughly three times Retell's support stack.
CloudTalk's appeal is the handoff. When the AI agent can't help, the call lands with a human on the same system, which answers Synthflow reviewers who couldn't reach anyone. Its Trustpilot score (4.1 from 823) is for the phone system as a whole. Choose it if you need seats for human agents anyway. We don't have a CloudTalk guide.
Goodcall
Goodcall prices the way a small business thinks: per agent, per unique caller, with unlimited minutes. Its pricing page lists Starter at $79 a month for 100 unique customers, Growth at $129 for 250 and Scale at $249 for 500, with $0.50 for each extra unique caller and 15% off on annual billing ($66, $108 and $208). A unique caller is "any caller with a unique phone number that calls and interacts with your Goodcall agent in any given month."
For a shop or clinic answering a few hundred calls, that is a fraction of Synthflow's floor. The unit rewards long calls and repeat callers, since neither costs extra. It punishes volume from new callers: at 2,500 first-time callers a month the Scale plan reaches $1,249. Trustpilot shows 2.5 from 5 reviews.
Choose it if you're a single-location business that wants a receptionist, not a platform. Our Goodcall guide goes further.
PolyAI
PolyAI's team designs, builds and tunes the agent with you, which makes it the managed option on this list. Its pricing page says use is "priced on a per-minute basis, which includes proactive performance improvements, maintenance and 24/7 support," without a rate. PolyAI does publish one on AWS Marketplace: a 12-month contract with a "minimum annual commitment for 500,000 minutes of usage" at $175,000, which is $0.35 a minute and no refunds.
That makes PolyAI more expensive to enter than Synthflow, not less: $14,583 a month against $2,500. It belongs here because it answers the support complaint in the most complete way, with 24/7 support and ongoing tuning written into the price. Trustpilot has only 2 reviews.
Choose it if you run a contact center with steady volume above about 42,000 minutes a month and want a vendor accountable for the agent's performance. Our PolyAI guide covers the product.
Also considered
Parloa and Cognigy are enterprise contact-center platforms sold by contract, like PolyAI. Replicant is a managed contact-center vendor. Voiceflow is a builder for chat and voice agents that you'd pair with your own telephony. My AI Front Desk is a small-business receptionist like Goodcall. Gartner's Synthflow alternatives page leans enterprise: Kore.ai, Cognigy, IBM watsonx Orchestrate, Sprinklr, Yellow.ai and Aisera. For the wider market, our guide to conversational AI platforms compares them by type.
What does each billing model reward?
The billing unit tells you what a vendor is optimizing for, and whose problem a slow month becomes.
| Billing unit | Who uses it | What it rewards | Who carries the risk |
|---|---|---|---|
| Annual contract floor | Synthflow, PolyAI | The vendor earns the same whether you use 5,000 minutes or 30,000 | You, until your volume passes the floor |
| Per minute, by component | Retell, Vapi, ElevenLabs | Longer calls earn the vendor more, and so do heavier models | Shared: you control length and model |
| Plan fee + per minute | Bland | Moving you to a bigger plan as daily call caps bite | You, at the cap |
| Flat plan, unlimited minutes | Thoughtly | The vendor wants short calls and caps concurrency | The vendor, up to its concurrency limit |
| Per unique caller | Goodcall | Repeat callers and long calls are free; new callers aren't | You, if you grow by new callers |
| Minute bundles + phone seats | CloudTalk | Selling seats for human agents alongside the AI | You, if you buy a bigger bundle than you use |
Synthflow's move to a floor makes sense for Synthflow: its docs now put account teams, launch support and contracted SLAs on every deal, and those costs don't shrink for a 1,000-minute customer. That is also why a small team gets better value elsewhere.
What moves when you switch from Synthflow?
Agent logic has to be rebuilt. Synthflow's agent docs describe "Import from JSON" for new agents but document no export, and they note that switching between Flow Designer and Prompt Builder means you'll "need to rebuild the agent in the other mode." Copy your prompts, node logic, variables and action definitions by hand before you cancel. Nothing on the other side reads a Synthflow flow.
The example in Synthflow's own docs shows what that copy job covers even for a simple booking agent: a global prompt, a greeting, conversation nodes that collect variables, a custom action calling an endpoint, two branches and a subflow. Each node's text and each variable has to be re-created in the new builder.
Knowledge bases move as source files. Re-upload the PDFs, documents and URLs you attached, or re-point the Zendesk Help Center import. Every alternative here takes documents or URLs.
Phone numbers depend on who owns the carrier. Synthflow's phone number docs say numbers you buy or import "belong to your workspace," and describe three routes: buying through Synthflow, importing over SIP, or connecting your own Twilio account. If your numbers live in your own Twilio or SIP trunk, you re-point them at the new platform. If you bought them inside Synthflow, the docs don't describe porting out, so ask Synthflow before you give notice.
Call history exports as CSV. Synthflow's export docs cover CSV exports and daily exports to Amazon S3, Google BigQuery or SFTP. Take a full export before the account closes.
Integrations mostly need re-wiring. Webhooks and custom actions are plain HTTP, so the receiving endpoints keep working once you point the new platform at them. Native apps, such as Synthflow's Freshworks and GoHighLevel integrations, don't carry over.
Contract timing. On an Enterprise agreement, check the notice period and any unused-minute carryover before you commit to a switch date. Run the new agent on a second number for two weeks before you move the main one.
What if your queue is tickets rather than calls?
Some teams land on this page because Synthflow was answering questions that could have been emails or chats. If most of your volume sits in a help desk, a voice platform is the wrong tool. Macha is an AI agent layer on the help desk you already use (Zendesk, Freshdesk, Gorgias, Front, HubSpot or Intercom): its agents read and reply inside the ticket, and billing is per ticket, from $299 a month for 750 tickets, with setup and monitoring by the Macha team included. It fits teams already running one of those help desks; it doesn't answer phone calls. The tiers are on our pricing page, and our guide to AI agents for customer service covers where voice and ticket agents each fit.
How we researched this
On September 19, 2026 we read the pricing pages of Synthflow and all eight alternatives, Synthflow's billing, usage, agent, phone-number and export docs, PolyAI's and Vapi's AWS Marketplace listings, and Trustpilot pages for every vendor. The Retell, Vapi, Bland, ElevenLabs and Synthflow prices match what our pricing posts on those vendors recorded the same day. We checked Gartner Peer Insights, Capterra and G2 in Chrome (Synthflow 4.5 from 1,015, Bland 5 from 11, checked September 19, 2026); the Reddit thread that ranks for this search wouldn't load. We did not build an agent on any of these platforms or create any vendor accounts, so every cost here comes from published rates and our stated assumptions. The two product screenshots, Synthflow's Flow Designer and Retell's agent editor, come from each vendor's public documentation, not from accounts of our own.
Frequently asked questions
What is the best alternative to Synthflow? For most small and mid-size teams, Retell AI: it is self-serve, pay as you go, includes 20 concurrent calls, and costs about $1,144 at 10,000 minutes a month against Synthflow's $2,500 floor. Developers who want to choose each component should look at Vapi; teams that want a managed service should look at PolyAI.
Is there a free Synthflow alternative? Several let you start free. Retell gives $10 of credit, Vapi $5, ElevenLabs Agents has a free plan with 15 minutes, and Bland's Start plan has no platform fee. None stays free at production volume; you pay per minute once the credit runs out.
Why did Synthflow stop offering monthly plans? Its docs say it now "uses sales-led pricing" and that the Pro, Growth and Agency plans "are no longer available for new subscriptions." New contracts start at $30,000 a year. Existing customers on the old plans keep them.
Which Synthflow alternative is cheapest? At 10,000 minutes a month, Thoughtly's Flex plan starts at $500 with unlimited minutes, if its 10-call concurrency fits. Among per-minute platforms, Vapi's calculator starts at $819 with its own telephony included, and ElevenLabs Agents comes to about $812 before telephony, roughly $898 with a Twilio local number.
Is Retell better than Synthflow? On price and self-serve access, yes for most teams under about 31,000 minutes a month. Retell has a 4.9 Trustpilot score from 814 reviews against Synthflow's 4.1 from 215. Synthflow's metering is more generous in places: seconds pool across calls, unanswered and voicemail-hangup calls bill a flat 5 seconds, and failed or user-canceled calls are free. That narrows the gap at high volume.
Can I move my phone numbers from Synthflow? If the numbers live in your own Twilio account or SIP trunk, you point them at the new platform. For numbers bought inside Synthflow, its docs don't describe porting out, so confirm with Synthflow before canceling.
Is Synthflow on AWS Marketplace? No. We found no Synthflow listing on AWS, Azure or Google Cloud Marketplace on September 19, 2026. Vapi and PolyAI both have AWS listings.
What is the best conversational AI platform for customer support? It depends on the channel. For phone calls, the platforms on this page. For email and chat tickets in a help desk, an agent layer that works inside the ticket is a better fit. Our conversational AI platform guide compares both.
If your support volume sits in a help desk rather than on the phone, you can start a Macha trial with $50 of free usage and no card, or see our tiers on the pricing page.
Sources: Synthflow pricing · Synthflow: Billing · Synthflow: Usage · Synthflow: Concurrency · Synthflow: Create an agent · Synthflow: Custom phone numbers · Synthflow: Export call data · Synthflow: Flow Designer · Synthflow blog: Voice AI cost (Aug 2025) · Capterra: Synthflow AI Receptionist · Trustpilot: Synthflow AI · Gartner Peer Insights: Synthflow AI alternatives · Retell pricing · Retell: Build your first phone agent · Vapi pricing · Vapi on AWS Marketplace · Vendr: Vapi · Bland pricing · Bland billing docs · ElevenLabs Agents pricing · ElevenLabs: Pay As You Go · Twilio US voice pricing · Thoughtly pricing · CloudTalk pricing · Goodcall pricing · PolyAI pricing · PolyAI on AWS Marketplace · Trustpilot: Retell AI · Trustpilot: Vapi · Trustpilot: CloudTalk · Trustpilot: Goodcall
Simple per-ticket AI pricing
One thread with one person is one charge, however many replies it takes.

