Overall MVP readiness stands at 40%. The nine components on the Build Status page show where each one is: Core Platform at 70%, Identity and Model Routing at 65%, Agent Runtime at 60%, Voice at 50%, Unified Workspace at 45%, Tools & Connectors at 40%, and the Developer Platform at 25% — architecture direction defined, public API not yet released. In Q4 the priorities are voice prototypes, durable workflows, and the connector foundation. The public website is also being rebuilt to match the product stage: one honest status, no absolute claims, and a clear distinction between working prototype, internal testing, integration, and planned. We will keep publishing here as milestones land — subscribe to DRENTIR Updates to get each one by email.
Development Roadmap: Q4 2026 and Beyond
DRENTIR is a working prototype, not a product — and the fastest way to make it real is to put it against real work. The Private Pilot is a limited program for a small number of teams: you bring one recurring workflow, we configure the prototype around it together, and what we learn shapes the roadmap. Pilot members get direct access to the team building DRENTIR and a vote on what gets built next — which connectors, which deployment options, which permissions come first. There is no fee and no SLA; this is joint development, not a purchase. If you have a workflow worth delegating, apply on the Business page.
Model Routing & Local AI is a working prototype at 65% readiness. The principle is simple: your hardware does the everyday work, and expensive cloud models are called only when the task truly needs them. Working now: local model execution, including GPU, and the routing foundation that sends each subtask to the right model. Every routing decision weighs privacy, latency, cost, capability, and available compute — inside the data boundaries you set. Still ahead: load-aware scheduling, benchmarking across models, and production reliability. The Private AI page documents the four deployment modes: local-only, hybrid, cloud, and private infrastructure.
The Agent Runtime — the layer that takes a plan and actually executes it — is now in internal testing at 60% readiness. Working now: the agent execution foundation and task orchestration across specialized agents. Still ahead: durable execution (tasks that survive restarts), failure recovery, and broader tool integration. Internal testing means the runtime is exercised daily against real tasks inside the development environment. It is not a public beta, and nothing here is presented as production software. When the remaining work lands, the runtime moves from internal testing to a working prototype — and the Build Status page will say so.
Q3 was about the foundation. The Core Platform — request intake, the planning foundation, and the agent execution layer — reached 70% readiness and is now a working prototype in internal testing. Identity, permissions and multi-tenancy reached 65%: permission levels and approval flows work, and the policy builder UI is next. A note on honesty: the percentages we publish reflect integration and validation, not lines of code. A component counts as done only when it is wired into the whole and tested there. In Q4 the focus shifts to the Agent Runtime, voice prototypes, and the first durable workflows. There are no launch dates yet — dates appear on the Build Status page only when the plan behind them is confirmed.