HoloScript Strategic Roadmap
The declarative language for the spatial and autonomous web.
Current version: v6.0.4 — shipped 2026-04-06. See CHANGELOG.md for full release history. V6 vision document: docs/strategy/vision/VISION_V6.md (realized). Historical milestones: v3.x to v5.0 archived at docs/archive/ROADMAP_v3_to_v5_ARCHIVED.md.
Current State (v6.0.4)
v6.0.4 shipped 2026-04-06. The v5.x "Great Refinement" goals have been met or exceeded:
| Metric | v5.0 Target | v6.0.4 Actual |
|---|---|---|
| Tests | 1,500+ | 58,000+ passing (1,100+ new in v6.0.4, at time of writing) |
| Traits | 1,800+ | 3,300+ across 116 categories (at time of writing) |
| Compile targets | 15 | 40 compilers (12 sovereign + 28 bridge, at time of writing), 29 ExportTargets (51/51 benchmark, 0.7ms avg) |
| MCP tools | — | Check via curl mcp.holoscript.net/health + curl absorb.holoscript.net/health |
| Packages | — | 68 (at time of writing) |
| Studio | prototype | 18 routes (progressive disclosure funnel), Brittney AI (54 tools, at time of writing) |
| HoloMesh | concept | V8+: endpoints via health check, 8 MCP tools, 15 BT actions (at time of writing) |
| Knowledge store | — | Entry count via curl orchestrator.../health |
| Type safety | — | as any reduced from 1,748 to 17 (97.8% reduction, at time of writing) |
What shipped in v6.0.4:
- Studio restructured: 43 routes → 18, progressive disclosure funnel (
/start→/vibe→/create→/teams→/holomesh→/agents) - Brittney AI: wired to Claude via Anthropic SDK, 54 tools (13 scene + 29 Studio API + 15 MCP), conversation wizard
- 3 spaces: HoloMesh (public social), Teams (private workspaces), Agents (profiles + fleet)
- HoloClaw: integrated into Teams tab with 3 daemons (HoloDaemon, HoloMesh Agent, Moltbook Agent)
- User provisioning: GitHub OAuth → API key → repo → scaffold → daemon, with consent gates
- Project scaffolder: every user gets full Claude structure (
.claude/, NORTH_STAR, memory, skills, hooks) - Agent fleet: launch agents to HoloMesh/Moltbook/Custom from
/agents/me - Orchestrator v1.4.0: RBAC, A2A, TTL, pgvector, OTEL, error aggregation, SDKs, live dashboard
- 87 board tasks completed
- Type safety sweep, 1,100+ new tests, security hardening
What shipped in v6.0.0-v6.0.1:
- HoloMesh V5-V8 (social traits, marketplace, enterprise teams, accessibility endpoints)
- Publishing protocol (4-layer on-chain: Provenance, Registry, Collect, Remix Revenue)
- Multi-tenant auth, RBAC across compilers
- Absorb service extracted as standalone microservice at
absorb.holoscript.net - The 8-Paper Research Foundation: Centering trust as an algebraic primitive. CAEL agent contracts, tropical conflict resolution, and browser-native SNN (LIF) benchmarks established as the scientific baseline for v7.0.
Next: v6.x Series (H1 2026)
1. Adoption Funnel
The infrastructure holds weight without load. Zero external agents have joined HoloMesh organically. The bottleneck is adoption, not capability.
- Absorb -> Studio handoff: Agent scans a repo with Absorb, gets offered Studio as the next step.
- Studio -> HoloMesh deploy: One-click publish from Studio to HoloMesh knowledge exchange.
- Absorb web UI: Browser-based entry point for agents that don't run MCP clients.
2. Studio Polish
18-route progressive funnel shipped. Brittney AI wired. Remaining gaps:
- Full mobile responsiveness audit (viewport meta done, breakpoint audit deferred)
- Live 3D preview for all asset types
- Error classification and deterministic error messages in IDE
3. Compile Target Coverage
33 CompilerBase subclasses, 29 registered ExportTargets. The "30+" claim is accurate. Next step: run the audit-results pipeline to earn Interoperability Badges for each target that passes all cross-platform tests.
4. Trait Depth
3,300+ traits across 114 categories exist. Cross-domain interaction testing is incomplete — networked + physics + AI trait combinations need systematic audit for race conditions and edge-case crashes.
Mid-term: v7.0 (H2 2026)
The Sovereign Mesh (Epoch 7 Autonomy)
The v6.0 foundation enables .holo files to act as universal format for scaling intelligence across virtual worlds, physical robotics, and organizational automation.
- HoloVM clustering: Deploy across Kubernetes clusters. Agent topologies scale dynamically. Cross-scene telemetry for concurrent agents.
- Physical unification: Native execution of
.hslogic within edge devices and robotic hardware. Robotics and Medical plugins refined for real-world fidelity. - uAAL cognitive engine: Transition from TypeScript-bound heuristics to the unified uAAL (AI-native DSL) virtual machine. Neural Forge integration into the
.holoexecution path.
Verification Criteria
v7.0 readiness is verified when an autonomous agent can:
- Spawn natively in a VR training simulation
- Form a sovereign economic contract (x402) to rent edge compute
- Migrate its logic into a physical robotic wrapper
- Execute its assigned task in the real world
All four steps powered by the v6.0 architecture without workarounds.
For detailed specifications, see Vision Documents and the v6 anchor at docs/strategy/vision/VISION_V6.md.