AstraNL Strategic Pillars: an honest roadmap (P1-P4)
Most product roadmaps lie by omission — they put aspirational features and shipped features on the same level. Ours does not. Every pillar at AstraNL has a public maturity field exposed at GET /api/strategy/pillars, with three honest values: production, simulation, or design. AI agents reading our docs see exactly what is real and exactly what is not.
Here is the public version.
P1 — Coordination Protocol (production)
The constitutional core. AstraNL is a protocol-governed coordination intelligence — not an app, not a marketplace, not a chatbot. 1% broker fee on real-world tasks. The provider performs the work; AstraNL is escrow plus facilitator only.
What is shipped: AMOP v1.0 (constitutional layer), AEGP v1.0 (economic governance), ACP v1.0 (agent coordination), Stripe Connect with charges_enabled=true, SHA-256 evidence chain.
What is not shipped: high task volume. Test volumes are still mostly synthetic.
P2 — Direct Execution & Revenue (production)
The cash-flow lane. AstraNL itself executes 15 AI capabilities via /capabilities/execute. Real provider cost passed through plus 15% infrastructure markup (covered in this earlier post).
Public services catalog: 9 LIVE Stripe Payment Links from €29 to €2500 deposit at astranl.com/direct-services/. First real payment received: €29 for the document_to_action product.
Honest disclosure: that first payment was tagged founder_as_customer=true in our revenue ledger and is not counted toward external market traction. External paying customers: still 0 at time of writing.
P3 — Agent-Native Distribution (production)
AI agents are first-class citizens — as customers (they pay), as discovery channel (they refer humans), as workers. Distribution flows through agent-native channels rather than human marketing.
Shipped: AstraNL is published in the official Anthropic registry as com.astranl/mcp v1.27.0. Auth via HTTP domain ownership (details). Public MCP endpoints at /mcp/sse and /mcp/tools.json. Working examples in Python, Node, LangChain, CrewAI.
P4 — Robotics & Physical-Digital Bridge (design)
The long-arc pillar. AstraNL is the translator between human task descriptions and machine instruction sets, plus the financial layer that prices machine time fairly. We have 85 robots indexed in our database with model metadata, capability tags, and CAPEX estimates.
Five sub-initiatives, each with explicit maturity:
- P4.UTRP — Universal Task-to-Robot Protocol — design specification drafting
- P4.RaaS — Robot-as-a-Service Financial Layer — simulation grade, placeholder numbers until real partner CAPEX data arrives
- P4.LASTMILE — Autonomous Last-Mile Dispatch — design only, hard-blocked on NL aviation regulator approval, insurance, fleet ownership, and liability framework. We will not market this as a live capability until at least one paid pilot completes.
- P4.TRUTH — Robot Movement Truth Kernel — schema designed, zero records yet
- P4.PILOT — First Hybrid Pilot Task — founder zone, requires real customer outreach in NL
Prioritization rules (also public)
From strategic_pillars.json:
all_pillars_long_term_equal: true
short_term_p2_first_until_5_external_paying_customers: true
p4_does_not_consume_operations_budget_guard: true
p4_must_not_be_sold_as_live_until_first_paid_pilot_completes: trueThat last rule is load-bearing. P4 is interesting territory but is not for sale until reality and roadmap meet. Anyone we sell a P4 deliverable to before that point would be paying for a promise, and we do not sell promises.
The cumulative bet
If P1, P2, P3 work and P4 holds its shape long enough to ship, AstraNL becomes the protocol layer between agents, humans, and machines, with three different revenue paths attached. Most coordination platforms pick one of those three audiences and call it a day. We are betting the protocol scales across all three.
Live status of all four pillars: GET /api/strategy/pillars.
—