Synthflow

What is Synthflow?

Synthflow is an AI voice intelligence platform designed for enterprise telephone automation. Companies can use this technology to have AI handle incoming calls or make outbound calls, thereby providing customer support, filtering leads, scheduling appointments, sending reminders, verifying identities, and transferring calls to human operators when necessary.

The platform offers a visual Flow Designer, a knowledge base, integration with CRM and calendars, APIs, as well as telephone infrastructure. The official website currently relies on sales-driven enterprise contracts; the old version of the self-service packages is no longer used as the standard pricing for new customers.

Main functions of Synthflow

• Incoming and outgoing calls:Answer customer calls around the clock or make calls proactively based on a list.
• Natural conversation:Handles interruptions, pauses, overlapping voices, voicemails, and end intents.
• Flow Designer：Use visual nodes to design branches, conditions, transitions, and business actions.
• Appointments and Reminders:Read the calendar, create appointments and send confirmations, or reduce no-shows.
• Clue filtering:Collect requirements, budget, and qualification information, then enter them into the CRM.
• Knowledge base Q&A:Use the company’s information to answer common questions and adhere to the prescribed wording.
• Manual transfer:Route calls to the appropriate team based on intent, risk, or business rules.
• Logging and analysis:View transcription, recordings, results, errors, and batch export.
• White-labeling and sub-accounts:Provides branded voice capabilities for SaaS, CRM, and service providers.

Phone scenarios suitable for automation

Synthflow price

The official corporate contract starts at $30,000 per year; the final price is determined based on the volume of calls, concurrent connections, phone plans, integration options, security measures, and support services provided. For new projects, it is necessary to contact sales for an assessment – the monthly fees that were previously made public should not be used as a standard for current contracts.

The existing billing arrangements for accounts that already exist will continue to be supported, and they can be viewed in Workspace Settings. If an increase in capacity or adjustments to the terms are needed, it is necessary to contact the customer support team; it cannot be assumed that the previous arrangements will still apply to new users.

Comparison of telephone connection methods

Tutorial on creating AI phone agents

1. Select a telephone scenario with clear boundaries, quantifiable parameters, and low risk as a pilot project.
2. Organize business objectives, the knowledge that can be answered, what commitments are prohibited, and the conditions for manual escalation.
3. Use Copilot or Flow Designer to create start, identity verification, branching, and end processes.
4. Connect to test numbers, calendars, or CRM sandboxes; do not write data to the actual customer systems from the start.
5. Tests for interruptions, silence, accents, noise, voice mail, rejections, and unexpected issues.
6. After meeting the acceptance criteria, the traffic volume is increased gradually, along with ongoing sampling for re-evaluation and optimization.

Flow Designer and Business Action Tutorials

1. Break the dialogue down into five categories of nodes: objectives, information collection, judgment, actions, and exit.
2. Set validation rules for each field to prevent the AI from entering vague responses directly into the system.
3. Add reservation, SMS, notification, Webhook, API, or manual transfer actions.
4. Design fallback paths for design timeouts, interface failures, lack of available time, and customer rejections.
5. Use test calls to cover all branches, and check the parameters and errors in the logs.
6. Set rates, concurrency, budgets, and alerts in the production environment to prevent abnormal bulk dialing.

Tutorial for launching outbound call campaigns

1. Confirm the source of the list, the approval status, the allowed calling times, as well as the requirements for blocking calls and unsubscribing.
2. Clean the phone number and set the region, time zone, caller ID, and maximum number of attempts.
3. Assign a clear identity to AI, and avoid leading users to believe that it must be a human.
4. After running in small batches, check the connection rate, goal achievement rate, complaints, and the quality of manual transfers.
5. Manual approval is required for sensitive decisions, so that AI cannot make decisions related to credit, healthcare, or insurance on its own.
6. Respect requests to cease contact, and update the CRM and block lists promptly.

Integration and development capabilities

• CRM, calendar, contact center, Webhook, and REST API;
• Twilio, Telnyx, RingCentral, Vonage, and the SIP/PBX area;
• Enterprise phone environments such as Asterisk, Cisco, Avaya, Genesys, etc.;
• Custom business actions, knowledge sources, and embedded components;
• Management of white-label sub-accounts, templates, permissions, and usage levels;
• The official GitHub organization provides various integration projects such as n8n nodes and skills.

Which users are it suitable for

• Customer Service Center:Handle routine high-frequency inquiries, as well as calls that come in during peak hours or outside regular working hours.
• Sales Team:Filter leads, schedule demos, and follow up on activities.
• Medical institutions:Handle administrative appointments and reminders within the framework of compliance.
• Financial and insurance institutions:Automated low-risk queries and document processing.
• CRM and SaaS platforms:Provide Voice AI within products under a white-label model.
• Large enterprises:SIP, regional deployment, concurrency planning, and strict SLAs are required.

Product advantages

• Integrate voice models, telephony, processes, business actions, and analytics into a comprehensive platform;
• The visual Flow Designer facilitates joint review by business and technical teams;
• Supports incoming calls, outgoing calls, scheduling, forwarding, and various enterprise integrations;
• It offers white-labeling and APIs, making it suitable for integration into platform-based products;
• Supports regional infrastructure, SIP, and high-concurrency enterprise deployments;
• It possesses compliance capabilities for SOC 2, HIPAA, and GDPR.

Usage restrictions and precautions

• The annual contract starts at $30,000, which is not suitable for users who only need a few personal phone calls;
• The total cost includes not only the platform fee but also calls, numbers, implementation, integration, and support;
• AI may mishear names, numbers, negative words, and accents; critical fields must be verified.
• Outbound calls are subject to regulations regarding regional communications, automatic dialing, marketing consent, and caller ID.
• Medical, financial, insurance, and identity-related applications require additional compliance measures as well as human oversight.
• Failed real-time integration can lead to duplicate appointments, incorrect updates, or incomplete transfers.
• The built-in Stripe reselling feature within the white-label option will cease to be available on September 15, 2026, and migration is required.

Security, recording, and data control

Synthflow states explicitly that it complies with the requirements of SOC 2, HIPAA, and GDPR, and offers options for enabling or disabling recording and transcription. Companies can develop strategies based on factors such as regional deployment, permissions, logging, and data retention policies.

• Before recording, the local requirements regarding mutual or unilateral consent must be met;
• In sensitive scenarios, recording or transcription can be disabled, but this will affect the quality inspection and analysis capabilities.
• Using Azure OpenAI can meet some of the data processing requirements in the European Union;
• APIs, Webhooks, and CRM credentials should be granted the minimum level of access necessary, and they should be rotated regularly;
• Before going live, DPA tests, security reviews, and data deletion process tests must be completed.

API and open-source information

Synthflow offers API, Webhook, custom actions, and white-label embedding capabilities; the specific rates, concurrency levels, and functions are determined by the enterprise contract. Some n8n nodes and Agent Skills projects are available on the official GitHub repository.

The core voice AI agent platform, telephone infrastructure, and hosting services are not open-source products. The publicly available integration repositories do not mean that the entire Synthflow system can be deployed on one’s own.

Frequently Asked Questions

How much is Synthflow?

The enterprise contracts available on the official website start at $30,000 per year; the final price depends on factors such as minutes used, concurrent connections, phone calls, integration needs, security requirements, and support services after deployment. The prices listed for the old self-service version should not be used as a basis for pricing new customers.

Can Synthflow be connected to existing phone numbers?

Existing phone systems can be connected using Twilio’s or Enterprise’s SIP/PBX solutions, or Synthflow numbers can be purchased. The specific rules regarding migration and caller ID must be verified depending on the region.

Does Synthflow support outbound calls?

It supports incoming and outgoing calls, and can be used for lead filtering, scheduling, and reminders. Companies must comply with regulations regarding automatic dialing, marketing consent, cancellation of subscriptions, and allowed calling hours.

Is Synthflow suitable for medical institutions?

The platform offers HIPAA-compliant capabilities that can be used for administrative tasks such as scheduling appointments, but it is necessary to sign the relevant agreements and impose restrictions on the data. AI cannot replace medical diagnosis, treatment, or emergency services.

Is Synthflow open source?

No, the core platform is designed to provide business services for enterprises. The integration components and features available on the official GitHub do not constitute the complete source code of the voice platform.