Agent Network Protocol (ANP)
An open protocol for agent communication, defining identity, discovery, negotiation, and secure messaging for the Agentic Web.
- Source repo
- agent-network-protocol/AgentNetworkProtocol
- Stars
- ★ 1.5k
- Last updated
- 4d ago
- License
- Apache-2.0
- Primary language
- HTML
- FA score
- 31/100 · Major gaps
At a glance
- How it runs
- Works with
- Universal · cross-platform
- Cost
- Free, no paid service needed
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- Developers building interoperable agents that need to authenticate, discover, and communicate across different platforms.
- Not a fit if
- Teams needing a production-ready, stable SDK out of the box
- Developers wanting one document to implement the protocol
- Projects requiring a finalized meta-protocol or stable payment standard
- Source review
- 31/100 · Major gaps 8 safety controls not found
What does this agent do, and when should you use it?
Agent Network Protocol (ANP) is an open-source protocol suite designed to become the HTTP of the Agentic Web era. It builds on existing internet infrastructure (HTTP, CA, DNS, CDN, TLS) and defines standards for agent identity (did:wba), naming (WNS handles), agent description, agent discovery, end-to-end encrypted instant messaging, and the AP2 agent payment protocol. The repository primarily contains specification documents, organized into a technical white paper and a series of numbered protocol documents (ANP-03 through ANP-10), along with a messaging profile suite (P1-P9) and draft vNext extensions for multi-device support. The actual SDK implementation is maintained in a separate repository (AgentConnect). ANP emphasizes a protocol-centric approach over platform-centric lock-in, aiming to enable cross-platform, cross-domain agent collaboration.
ANP defines protocol specifications for agent identity authentication (did:wba), naming (WNS handles), agent description, discovery (active and passive), end-to-end encrypted messaging (via nine interoperable profiles), and agent payments (AP2). It provides indexed documentation, quick-start guides, sample programs, and example assets (e.g., ADP assets, API interface examples, vNext JSON examples). The protocol operates by defining JSON-RPC 2.0 bindings (P1), DID-based service discovery (P2), direct/group messaging semantics (P3/P4), E2EE overlays (P5/P6), attachments (P7), federation (P8), and mentions (P9). The meta-protocol (ANP-06) remains a draft and is not released. The project also offers a reference SDK (AgentConnect) for practical implementation.
- Developers building interoperable agents that need to authenticate, discover, and communicate across different platforms.
- Teams looking to implement DID-based identity and end-to-end encrypted messaging for their agent systems.
- Organizations seeking to avoid platform lock-in by adopting an open protocol for agent collaboration.
- Projects requiring agent-to-agent payment capabilities (AP2) or secure multi-device messaging (vNext draft).
- Architects or researchers studying agent interoperability standards and protocol design.
How do you install or deploy this agent?
ANP is a specification repository and does not require installation. To build on ANP, use the official SDK (AgentConnect):
git clone https://github.com/agent-network-protocol/AgentConnectFollow the SDK's documentation for language-specific setup and dependencies.
How do you use this agent?
To start, read the [ANP Getting Started Guide](docs/anp-getting-started-guide.md). For identity implementation, start with ANP-03 (did:wba) and the DID compatibility appendices. To publish an agent, read ANP-07 (Agent Description) and ANP-08 (Agent Discovery). For messaging, begin with ANP-09 and select the required messaging profiles (e.g., P3 Direct Messaging, P5 Direct E2EE). Refer to the [sample programs](docs/chinese/ANP示例程序.md) for runnable examples.
What are this agent's strengths and limitations?
- Grounds identity in W3C DID and reuses existing internet infrastructure (HTTP, TLS, DNS), reducing adoption friction.
- Provides a comprehensive protocol suite covering identity, naming, discovery, messaging, and payments.
- Clear versioning and compatibility guidance, including k1_ extension and did:web comparison, aids migration from existing systems.
- Messaging is split into nine interoperable profiles, allowing incremental adoption and customization.
- Active open-source community with dedicated communication channels and contribution guidelines.
- The meta-protocol (ANP-06) is still a draft, meaning some core negotiation capabilities are not finalized.
- The specification is extensive and requires deep reading to implement correctly, posing a learning curve.
- SDK implementations are not fully aligned with the 1.1 spec set, potentially causing compatibility issues.
- No official releases or stable version tags, so maturity and production-readiness are unproven.
- Relies on community contributions for maintenance, which may affect update frequency and long-term support.
How does this agent compare with similar options?
No direct competitors are mentioned in the source, so no comparison is provided.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Agent Network Protocol (ANP) This agent | 31 · Major gaps | Library / SDKFree | ★ 1.5k | 4d ago | HTML | — |
| Prismer Cloud | 56 · Major gaps | CLIFreemium | ★ 1.4k | 4d ago | TypeScript | Claude Code |
| Bindu Agent Infrastructure | 62 · Some gaps | Library / SDKFree + model costs | ★ 10k | 1mo ago | Python | OpenAI API · Claude API |
| CC-Connect | 60 · Some gaps | CLIFree + model costs | ★ 16k | 3d ago | Go | Codex · Claude Code |
How does FollowAgents rate this agent?
Why each dimension lost points
The evidence shows this is a protocol specification repository without executable code or runtime permission model. Therefore, least privilege, user confirmation, data flow transparency, sensitive data handling, dependency security, external effects, rollback, and source attribution cannot be assessed, scoring 0. The repository provides no implementation details regarding permissions or data handling.
Self-consistency: README and specification documents are consistent in version numbers, protocol status, and architecture description, scoring 2. Dependency availability: The repository has no runtime dependencies, but references an external SDK (AgentConnect) whose availability is unverified, scoring 1. Failure messages: No documentation of error handling or failure messages is provided, scoring 0.
Audience and scenarios: README clearly identifies target audience (developers) and use cases (agent communication), scoring 2. Capability boundaries: Protocol specifications define clear capability scope (identity, discovery, messaging, etc.), scoring 2. Trigger precision: No trigger mechanisms or event definitions are provided, scoring 0. Environment fit: Describes architecture based on existing internet infrastructure, but lacks specific deployment environment requirements, scoring 1.
Information architecture: Repository structure is clear, documentation is well-categorized, scoring 2. Install notes: No installation or usage instructions are provided, scoring 0. Naming stability: Protocol versions and naming conventions are clear, scoring 2. Examples and FAQ: Provides links to example programs, but no FAQ, scoring 1. Known limitations: Clearly indicates draft status of meta-protocol and unpublished content, scoring 2. License: Provides Apache-2.0 license, but README declares MIT, inconsistency, scoring 2. Versioning and changelog: Has version notes and milestones, but no formal changelog, scoring 2. Maintenance responsibility: Provides contact and contribution guidelines, but does not clearly identify maintainers, scoring 1.
Output usability: Protocol specifications can serve as implementation reference, but no directly usable output is provided, scoring 1. Marginal value: Protocol fills a gap in agent communication domain, high value, scoring 2. Cost-benefit: No performance or cost analysis provided, scoring 1.
Claim traceability: Some claims in README can be traced to specification documents, but some lack evidence, scoring 1. Cross-source corroboration: References external SDK and community, but no independent verification, scoring 1. Fact-inference separation: Document distinguishes published and draft status, but some content is visionary, scoring 1.
- Not found in source: least-privilege scopingGrant only what the task needs: a dedicated account or read-only token, scoped to specific directories and repos.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: data-flow disclosureWatch which external services it contacts (proxy or firewall logs) and keep sensitive data out until you know where it goes.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
- Not found in source: disclosed external effectsEstablish which external systems it writes to, sends to or changes, and verify with test accounts or repos before production.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- Not found in source: verifiable attributionInstall from the official repo or registry and check the publisher and URL to avoid look-alike packages.
- License inconsistency: README declares MIT, but LICENSE file is Apache-2.0, needs clarification.
- Repository is specification-only, no actual code, security and reliability cannot be assessed.
- Meta-protocol and some messaging profiles are still drafts, version compatibility should be considered during implementation.
Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
v1.2 ANP 1.2 Release
Release notes excerpt
ANP 1.2 Release ANP 1.2 is the current specification-set release. It promotes the previously frozen vNext drafts into their canonical document paths. This is a documentation release: it does not change SDK, service, or product implementations, and it does not authorize public capability advertisement. Highlights - ANP-02 as the common authentication entry Adds the released DID-method-independent HTTP/JSON authentication specification for did:wba and native did:web. Onboarding and reader entry now point to ANP-02, not ANP-03. - Promote core identity and naming specs …View on GitHub ↗ -
v1.1 ANP 1.1 Release
Release notes excerpt
ANP 1.1 Release ANP 1.1 is an important protocol-suite update focused on improving identity, naming, and agent-to-agent communication. Highlights - Improved identity authentication design Refined the did:wba identity and authentication mechanism to make cross-platform agent authentication clearer, more interoperable, and easier to implement. - Added WNS naming support Introduced the WNS Handle scheme as a human-readable naming layer for DID-based agents, including handle-to-DID resolution and DID rotation support. …View on GitHub ↗ -
V1.0 1.0
Release notes excerpt
🎉 ANP 1.0 Official Release We are proud to announce the official release of ANP 1.0! Building on the core capabilities of ANP — agent identity, agent description, and agent discovery — this release introduces significant enhancements in security, privacy, and trustworthy transactions, including: - ✅ Human-in-the-loop authorization for critical operations such as financial transactions, adding a strong layer of operational oversight - 🔐 Support for multiple DIDs and periodic DID rotation, enabling stronger privacy protection for agent users …View on GitHub ↗