Use right after top-up · Balance never expires · Live same day
Pay as you go, billed per second:
Compute from ¥0.05/min, calls from ¥0.1/min, billed by actual call duration.
Flexible billing options:
Pay as you go, pay per result (¥8 per qualified lead, ¥18 per WeChat add) or monthly seats (from ¥599).
Top up from ¥2000, no minimum spend:
Top up from ¥2000; balance never expires, even if unused for a long time.
Sign up and go live the same day:
New users get trial credit on sign-up and can start calling the same day the script is configured.
Own your data · white-label resale
Standalone full-system deployment:
Frontend, backend, database and admin all run on your own servers — a fully dedicated system.
Own your data, unlimited concurrency:
Call records and customer info are stored in your own database, with no platform limit on concurrency.
White-label ready for resale:
Replace the logo, domain and brand colors in the admin console to serve customers under your own brand.
Delivered in as fast as 7 business days:
Standardized modular deployment, about 10x faster than on-premises.
Fully local data · strict compliance
All AI compute runs on-premises:
Speech recognition, synthesis and LLM all run on local GPUs; data never leaves your network.
Meets compliance requirements such as MLPS Level 3:
Apart from the carrier voice channel, no data leaves your systems — suited to strict-compliance settings like healthcare and government.
Project-based delivery with dedicated implementation:
Full implementation services: environment setup, performance tuning, load testing and security hardening.
Deep customization for large organizations:
For top-tier hospitals, governments and conglomerates with the strictest data security requirements.
Beyond call charges, these modules are billed only when used. Nothing to pay if you don't turn them on, and no extra plan to buy.
Clone a real voice from one recording. One-time fee, reusable across all your AI agents.
Collect business leads by industry keyword and region. Billed by leads collected, rounded up to the next 1,000.
Connect WeChat or WeCom accounts to an AI agent to handle chats and groups 24/7. Renews monthly.
Compute is billed by conversation time with no phone charge; qualified leads are billed at the per-result rate.
Standard script generation, training-center tuning and knowledge-base imports are free. Adjust any time.
Tailored to internal roles such as engineering, operations, documentation and finance, and deployed privately on your own servers or workstations so data never leaves your company. Every action that makes a change requires item-by-item approval from a manager.
Estimated cost
Per agent per day 800 calls
Current billing:Pay as you go; intent and WeChat adds are results fees, added on top.
Compute ¥0.05/min + call ¥0/min; qualified lead ¥8/each, add WeChat ¥18 each (results fees, not added to compute or call fees); monthly digital employee ¥599/month each, compute included, calls still billed at ¥0/min extra. Estimates only; the signed plan prevails.
Cloud: active on top-up, call the same day. Private cloud: 3–7 business days. On-premises: 6–8 weeks.
No charges for unconnected calls such as busy, powered off, invalid, suspended or no answer.