Building trust infrastructure for people, agents and applications.
AgentCredia started with payments for AI agents. Our direction now covers identity, authentication, verification, reputation and the evidence application owners need to understand human and agent interactions.
代理即将开始花钱。基础设施尚未准备好。
到 2025 年,AI 代理已经从令人印象深刻的演示跨越到真正有用的产品。它们在调用 API,阅读文档,编写代码,运行工作流。到 2026 年初,下一步变得无法忽视:这些代理需要支付费用。高级 API,付费 MCP 工具,其他代理,具有真实成本的真实服务。
现有选项都不适合问题的形状。卡网络不会为低于一美元的支付定价。Stripe要求人类的入职流程。自我保管钱包超出了大多数代理开发者想要操作的范围。即使是新兴的稳定币通道,尽管它们很好,但也没有提供代理平台所需的操作原语 - 每个代理的身份、每个代理的支出政策、每个代理的审计日志。
AgentCredia 的存在是为了填补这一空白。我们没有发明稳定币轨道(Circle 发明的),我们没有发明 x402(Coinbase 发明的),我们也没有发明代理(每个人都在争夺这个)。我们构建的是一个运营平台,位于代理开发者和这些轨道之间 - 开发者在这里创建钱包、附加身份、设置支出政策,并在十五分钟内获得一个可工作的有偿代理。
十八年来一直在同一个未来中发货。
AgentCredia 是创始人第三个创业项目,他在职业生涯的大部分时间里致力于构建早期基础设施。我们的创始人兼首席执行官 Toshendra Sharma 在印度理工学院孟买计算机科学毕业后不久创办了他的第一家公司。在 AgentCredia 之前的几年里,他在区块链和应用 AI 的交叉点上构建了另外两个已交付的产品 - 一个自筹资金实现盈利,一个获得风险投资 - 并因此工作被评选为福布斯亚洲 30 位 30 岁以下名单。
这两家公司以不同的方式教会了同样的教训:加密货币采用的瓶颈从来不是轨道本身,而是其上方的操作层。钱包不够;支付 API 不够;身份不够。团队需要的是一个将这些原语捆绑成适合他们实际构建的工作负载的开发者表面的平台。在 2018 年,该工作负载是代币化记录。在 2022 年,它是 AI 辅助教育。在 2026 年,它是需要花钱的自主 AI 代理。
The pattern of "everything you need on top of stablecoin rails, for one specific workload, with one opinionated surface" is what AgentCredia is the latest iteration of. The team has shipped this shape of product before, against earlier workloads, with earlier rails. The execution risk is well-understood; the bet on which workload to back is what is new.
Toshendra Sharma is also the founder of Proof of Actor. Credia Proof-of-Actor brings the application-facing Proof of Actor direction into the AgentCredia product family.
From a payment focus to a broader product family.
Credia Payments carries forward the payment offering. Identity, Auth, Verify and Reputation give the wider direction explicit names and responsibilities. Credia Proof-of-Actor brings the Proof of Actor direction into the family as an application-facing product. Each offering has its own availability and scope.
- Credia IdentityPrivate Beta. Available to selected teams. Request access to discuss your integration.Establish an agent's identity and its connection to an operator.
- Credia AuthPrivate Beta. Available to selected teams. Request access to discuss your integration.Define how an agent authenticates and what authority it can use.
- Credia VerifyPrivate Beta. Available to selected teams. Request access to discuss your integration.Inspect specific claims and the evidence behind them.
- Credia ReputationPrivate Beta. Available to selected teams. Request access to discuss your integration.Understand activity and outcomes in context over time.
- Credia PaymentsLive. Get started with Credia Payments.Payment workflows and controls for AI agents.
- Credia Proof-of-ActorPrivate Beta. Available to selected teams. Request access to discuss your integration.See if the traffic in your app is human, AI-assisted or agent traffic.
我们愿意接受评估的原则。
我们是一个位于印度的小团队,与美国、欧洲和亚洲的客户合作。产品发展迅速,因为团队精简;团队能够保持精简,因为产品有明确的方向。少数原则塑造了每个决策。
默认非托管
我们从不在平台层持有客户资金。钱包由代理人拥有;我们提供其上方的操作层。这既是安全保证,也是监管简化。
政策在 API 中,而不是在提示中
支出政策、白名单和时间窗口都存在于钱包API中,并在服务器端进行评估。将它们放在代理的提示或代码中会在提示注入下使其失效 - 而且提示注入是真实存在的。
定价透明,按代理计算
Free, Pro, Business. Per agent, not per workspace. You only pay for the agents you actually run. The fee schedule is published; no contract minimums for the self-serve tiers.
将无聊的部分做好
Webhook具有HMAC签名。审计日志在Business上具有哈希链篡改证据。SDK具有可工作代码,而不是伪代码。无聊的部分是信誉所在。
可运行的文档
我们文档中的每个代码片段都在CI中执行。如果集成步骤在午夜时对开发者不起作用,我们就不会发布它们。损坏的示例是故障单,而不是代码片段。
构建AgentCredia的人。
一个拥有深厚区块链、SaaS 和运营经验的小型创始团队。总部位于印度,客户遍布全球。

创始人兼首席技术官
Rohendra Singh
程序员、架构师、DevOps 负责人,拥有十五年已交付软件开发经验。曾在区块链和教育领域将七个以上安全 SaaS 产品投入生产环境。负责 AgentCredia payment API 和 chain-adapter 架构。

运营主管
Kanishka Sharma
运营和客户成功负责人,重点关注留存和客户体验。负责 onboarding、支持响应时间,以及推动路线图优先级的反馈闭环。擅长把支持中的模式转化为产品修复。

创意负责人
Sharjeel Maniyar
具备交互设计和前端开发经验的产品设计师。孟买大学 IT 毕业。负责仪表盘、公开代理资料页以及本网站所属营销页面的视觉与交互体验。
Two sides of every agent interaction.
On one side are the teams that build agents: AI agent builders, MCP server operators, API providers and agent platforms. They need identity, explicit authority, evidence and payment workflows for the agents they ship.
On the other side are the teams that operate the web apps, mobile apps and APIs those agents use. They need to understand whether a person, a person working with AI, or an agent is on the other end, and Credia Proof-of-Actor is built for them.
Explore solutionsTrust should be inspectable.
We want teams to understand which actor is involved, what authority it has and what evidence supports a claim. Product claims should be as clear as the interfaces developers use.
