Thorbase Inc.
Thorbase Inc. operates TokenGO, an enterprise-grade LLM API aggregation and inference platform. TokenGO gives developers and AI-native teams one OpenAI-compatible integration to access multiple leading models, with smart routing, cost efficiency, failover, and privacy-focused operations.
At a Glance
- AI-native businesses where token spend is a significant budget line
- Development teams running meaningful token volume
- Enterprises needing US contracting, governance, DPA, and service-level guarantees
- Teams building production AI agents and coding assistants
- +1 more
AI Tools by Thorbase Inc.
(1)
TokenGO
Unified AI Inference API Gateway
Discussions
No discussions yet
Be the first to start a discussion about Thorbase Inc.
Latest News
Compute prices could surge over 10x in the future
Agentic workloads are inherently burst-heavy
Infrastructure requirements for running agents in production
Kimi K3, the first open-source model in the 3-trillion parameter tier
Products & Services
A serverless, OpenAI-compatible API and model aggregation platform that provides one key and one integration surface for models from providers including OpenAI, Anthropic, Google, DeepSeek, Moonshot, Z.ai, MiniMax, and Qwen.
Web console for account management, Stripe balance top-ups, API-key creation, quotas and expiration, billing, model discovery, playground testing, and detailed usage logs.
Enterprise offering with 99.9%+ SLAs, automatic provider failover, reserved compute availability, team token assignment, spend controls, audit visibility, DPA/compliance review, custom billing, and migration support.
Market Position
TokenGO positions itself as a cost-efficient, production-oriented alternative to managing separate direct model-provider integrations: it combines a unified OpenAI-compatible gateway with its own routing, elastic scheduling, inference optimization, caching, and datacenter-partner supply. Its differentiators are published volume pricing, fallback routing, enterprise contracting and SLAs, and a stated zero-retention/no-training posture.
Leadership
Founders
Eddie Zheng
Founder of Thorbase; LinkedIn identifies his education as the University of Chicago and lists prior roles as a Private Equity Intern at SAIF Partners and Summer Analyst at ZheShang Securities Co. Ltd.
Executive Team
Eddie Zheng
Founder
Founder of Thorbase; LinkedIn lists University of Chicago education and prior internships at SAIF Partners and ZheShang Securities Co. Ltd.
Founding Story
Eddie Zheng said his team had been building TokenGO for about five months when he announced it on LinkedIn. The initial vision was to make access to AI inference more cost-effective by aggregating idle datacenter compute and routing requests to available supply, while giving customers a single API rather than requiring separate integrations for each model provider.
Business Model
Revenue Model
Pay-as-you-go usage priced per million input and output tokens, with monthly-volume discounts, quarterly billing for Growth, and custom enterprise billing. Customers can top up account balances through Stripe; enterprise contracts use USD wire/ACH invoicing.
Pricing Tiers
List pricing; unified model catalog, cost and usage monitoring, per-key limits, and unlimited seats.
20% volume discount; quarterly billing, priority support, SLA and professional services, team management, and access controls.
30% volume discount; dedicated support, DPA, and compliance review.
30%-50% volume discount; custom billing and invoicing, with enterprise reliability, governance, contracting, and reserved-compute options.
Target Markets
- AI-native businesses where token spend is a significant budget line
- Development teams running meaningful token volume
- Enterprises needing US contracting, governance, DPA, and service-level guarantees
- Teams building production AI agents and coding assistants
- Developers seeking a single API for leading open and closed-weight models
- Production LLM inference for AI-native businesses
- Agentic workloads with bursty, tool-using, multi-step execution
- Coding assistants and other high-volume developer applications
- Long-context analysis and applications using large-context models
- Teams that want lower inference costs without rewriting an OpenAI integration
- Organizations needing provider redundancy, failover, governance, and predictable capacity