Weave
Read SDK →
For autonomous software

Agents are
first-class readers.

The Weave protocols and docs are designed for autonomous software as a primary audience. Identity, capabilities, transport, and documentation are all consumable by humans and agents — without scraping, without guessing, without bespoke wrappers.

Context

/llms.txt

Compact directory of every navigable doc page, designed to fit a single LLM context.

Open
Full context

/llms-full.txt

Complete docs in markdown form, ready to load into AI tools wholesale.

Open
Capabilities

/skill.md

Capability summary for agents and skill systems — what Weave can do and how to use it.

Open

Agent capability map.

GoalPrimary libraryPair with
Locate peersweave-dhtweave-dht-daemon
Message peersweave-swarmweave-secret-stream
Share filesLocusAgentFS
Store eventsTraceTrace Blobs
Build social appsWOVENNexus, Lens
Sign and verifyweave-identityweave-crypto
Build a CLIdcliweave (binary)
Run a nodeweave-sdkweave-core

Constraints agents should assume.

  • • No global consensus, no chain finality, no token economics.
  • • Local verifiability is the substrate. If it isn't signed, it isn't trusted.
  • • Replication is best-effort; convergence is per-view via Nexus.
  • • Source repositories are authoritative for code-level details.
  • • Docs pages describe library boundaries; load /llms-full.txt for system-wide context.