Verified project record

GetBindu/Bindu

Bindu is an identity, communication, and payments layer for AI agents that exposes decentralized identifiers, OAuth2 authorization, mutual TLS, and USDC payment collection behind a single function call. It targets developers and AI engineers who want to integrate agent-to-agent communication and monetization without building custom plumbing.

Project overview

It ships three security layers (mTLS, OAuth2, DID signature) enabled by default and allows agents to demand USDC payments on EVM chains before executing work.

Project type: AI Agent · Infrastructure
Use cases: Automation
Deployment: Refer to project documentation
License: License pending

Best for

Developers and AI engineers building automated agent systems that require verifiable identity and payment enforcement.

Key capabilities

Exposes agents via A2A JSON-RPC supporting message/send, tasks/get, and message/stream.
Encrypts sockets and mutually authenticates them using X.509 certs from step-ca.
Signs agent responses with Ed25519 keys, allowing verification via W3C DID.
Validates caller authorization using scoped bearer tokens via Ory Hydra introspection.
Allows agents to demand USDC payments on EVM chains before executing work.
Enables agents to declare capabilities visible on their agent card.
Allows two agents to agree on price, latency, and SLA up front.
Exposes local agents to the internet via an FRP tunnel without port forwarding.

Limitations and risks

Requires Python 3.12 or higher.
Requires an API key for at least one LLM provider to run examples.
Has documented known issues that should be read before pushing to production.
Requires Postgres for tasks and messages storage, Ory Hydra for OAuth2, Smallstep step-ca for X.509 certs, and at least one LLM provider API.

Getting started

Setup is rated medium difficulty because it requires Python 3.12+ and the uv package manager, alongside coding a handler function and passing it to bindufy().

Alternatives and comparisons

An open-source super agent harness that researches, codes, and creates using sub-agents, memory, sandboxes, and extensible skills.
Provides a unified multi-provider LLM API, an agent runtime with tool calling, and an interactive coding agent CLI.
Provides autonomous AI penetration testing agents that dynamically run code, find vulnerabilities, and validate them.

Project comparisons

Evidence and sources

GitHub project description: Bindu: The identity, communication, and payments layer for AI agents.
README: You'll need Python 3.12+ and [uv](https://github.com/astral-sh/uv).
README: Bindu is all of that plumbing, behind one function call. You wrap your handler with `bindufy()`, and a few seconds later your agent is online with its own cryptographic identity
README: We're using Bindu in production to build the **Trade Compliance OS**
Release: 2026.22.4: x402 settle-first + scenario hardening

Reviewed at: 2026-07-16T12:25:37Z

Open original GitHub project