1.34K
total weekly downloads
92 packages
@weaveintel/a11y
Framework-agnostic accessibility utilities for weaveIntel UIs — pure, dependency-free helpers (focus restoration across full re-renders, etc.).
@weaveintel/a2a
A2A protocol — agent-to-agent communication client and server
@weaveintel/agents
Agent runtime — tool-calling loop, planner-executor, hierarchical supervisor-worker
@weaveintel/api-client
geneWeave typed API client — bearer-token /api/me surface over @weaveintel/client (runs, catalog, tasks, reminders, memories, devices, notifications, conversations). No React Native imports.
@weaveintel/artifacts
Generated shareable artifacts with redaction and sensitivity gating.
@weaveintel/cache
Semantic & operational caching — exact match, embedding similarity, tool result caching, TTL, invalidation
@weaveintel/capability-packs
Versioned, exportable bundles of DB rows ("capability packs") that can be installed and uninstalled atomically against any host that supplies a `PackInstallAdapter`. Phase 6 of the [DB-Driven Capability Plan](../../docs/DB_DRIVEN_CAPABILITY_PLAN.md).
@weaveintel/client
Browser-safe run client for the weaveIntel platform — SSE transport, progressive stream reducer, offline outbox.
@weaveintel/coedit
**CRDT co-editing for AI apps** — a tiny, zero-dependency toolkit so a human and an AI agent can edit ONE document at the same time and always converge, with awareness cursors and offline reconcile (Collaboration Phase 7).
@weaveintel/collab
Real-time collaboration + CRDT co-editing — one presence model, a CoeditDoc port with an RGA reference adapter, and the AI-as-editing-peer.
@weaveintel/collaboration
**Multiplayer primitives for AI apps** — shared sessions, presence ("who else is here"), live run subscriptions, and user handoff.
@weaveintel/compliance
@weaveintel/contracts
Completion contracts — task schemas, acceptance criteria, validation, evidence bundles
@weaveintel/core
Core contracts, types, and runtime for weaveIntel AI framework
@weaveintel/cost-governor
LLM spend budgets and limits — cap and govern model cost.
@weaveintel/devtools
Developer tools — scaffold, inspector, validator, mock runtime, migration
@weaveintel/durability
Durability primitives for the weaveIntel framework — idempotency, dead-letter queue, retry budgets, health, backpressure. Re-exports @weaveintel/reliability under the canonical durability name.
@weaveintel/encryption
Per-tenant envelope encryption: KMS providers, key manager, AEAD envelope, blind indexes
@weaveintel/equity-scoring
Deterministic, pure-function equity factor scoring engine with cross-sectional z-scores
@weaveintel/evals
Evaluation runner — assertions, scoring, suites
@weaveintel/extraction
Structured data extraction from text and documents.
@weaveintel/graph
@weaveintel/guardrails
Guardrail pipeline — evaluation chain, risk classification, cost guard, governance
@weaveintel/human-tasks
Human-in-the-loop orchestration — approval tasks, review queues, escalation, SLA enforcement
@weaveintel/i18n
Framework-agnostic internationalisation core for weaveIntel — pure, dependency-free message catalogs, an ICU-subset t() (interpolation + plurals), and BCP-47 locale fallback.