Rangoon site map

Start here

• Rangoon.ai — The control plane for governed AI

  A governed AI control plane for platform teams, government technology programs and development firms. Manage capabilities, policy, approvals, compatibility and execution evidence.

• Documentation

  Documentation explains Rangoon’s evolving capability model, authority boundaries, adapter contracts, operational evidence, and target-specific behavior.

• Open source

  Rangoon is designed as a free and open-source foundation for organizations to inspect, self-host, adapt, and contribute across capability, adapter, and governance layers.

• Roadmap

  Launch scope covers a durable capability model, reviewable adapters, governed operations, extension boundaries, and evidence-led delivery.

• AI documentation and public data access

  AI-ready documentation, open public data, feeds, and explicit permission for AI ingestion and training.

• Downloads & systems

  Deployment architecture for workstations, containers, servers, and bounded mobile roles. Review target responsibilities and release evidence; downloads open at launch.

• Marketplace preview

  A launch-preview catalog for skills, connectors, authority adapters, and enterprise extenders. Listings describe review requirements; none are installable or purchasable.

• Compatibility & ecosystem

  Explore agent runtimes, model providers, corporate software, data platforms and identity connections in Rangoon’s launch architecture.

• Rangoon site map

  A browsable index of Rangoon public pages, documentation, product material, and AI ecosystem reporting.

Platform

• Product

  Rangoon provides a control plane for composing, reviewing, testing, and preparing portable agent capabilities across changing harnesses and execution-authority systems.

• Agent fleet

  Design agent profiles with explicit roles, capability bundles, harness assignments, policy posture, and operational ownership before they enter a governed workflow.

• Skill registry

  Turn reusable procedures into inspectable, versioned capabilities with a declared purpose, inputs, dependencies, activation conditions, and target-aware compatibility.

• Workflows

  Model multi-step agent operations as readable graphs of inputs, decisions, tools, approvals, authority checks, outcomes, and evidence.

• Capability Studio

  Bring existing agent configuration into a structured lifecycle: discover, decompose, compose, test, compile, and prepare for governed deployment.

• Connectors

  Define explicit external-system operations so a connector’s inputs, credentials, side effects, policy context, and receipts can be reviewed as separate facts.

• Harness Manager

  Track runtime environments as evidence-bearing targets, then adapt capabilities deliberately instead of assuming that similar-looking agent tools behave the same way.

• Policy

  Make policy inputs visible where capabilities and operations are assembled, while keeping execution-time authority distinct from a policy result or runtime configuration.

• Test Lab

  Plan repeatable evaluation for capability activation, behavior, tool selection, output structure, policy handling, compatibility, and target-specific failure modes.

• Deployments

  Prepare exact, reviewable capability bundles with target evidence, dependencies, approvals, activation intent, and a recovery path before any environment changes.

• Analytics

  Plan operational analytics around execution outcomes, approvals, policy decisions, capability use, target health, latency, and evidence quality.

Architecture and developers

• Developers

  Rangoon defines reviewable contracts for skills, adapters, connectors, workflow nodes, authority integrations, tests, and interfaces before they reach a workspace.

• Harness adapters

  Build versioned boundaries that discover native configuration, normalize supported concepts, compile target artifacts, and explain exactly where a harness changes the capability.

• Connector contracts

  Plan connector contracts around explicit operations, data and side-effect metadata, credential needs, policy context, authority evaluation, and receipt behavior.

• LNSAT

  Rangoon uses LNSAT as its reference execution-authorization and evidence engine, separating capability management from a decision to perform a specific consequential action.

• Architecture

  Rangoon uses a canonical capability model, versioned target adapters, and an authority boundary that keeps configuration, deployment, and consequential execution from becoming the same operation.

• Standards and interoperability

  How Rangoon separates agent protocols, identity, policy evaluation, execution authority and operational evidence. Assess integration requirements against versioned release evidence.

• Security by explicit boundaries

  Rangoon’s security principles: configuration is not authority, connectors are not permission, and consequential actions require evidence.

Solutions

• Engineering

  Turn repository rules, test procedures, delivery checks, and coding-agent context into reviewable assets that remain explicit about each target’s behavior and limits.

• Security

  Define security-agent operations as scoped requests with declared capabilities, bounded connector actions, policy context, specific approvals, and execution evidence.

• Operations

  Plan operational agent workflows that make tools, systems, policy gates, human approvals, outcome states, and recovery signals visible to the people responsible.

• Enterprise

  Plan organization-wide capability management with environment hierarchy, delegated administration, shared libraries, policy lifecycle, evidence export, and rollout control.

• Domestic AI

  U.S.-first AI provider evaluation for technical teams, government and development firms. Review model selection, processing routes, permissions and execution evidence.

• Government

  Rangoon provides a launch architecture for teams that need to assess agent capabilities, workload boundaries, approvals, and operational evidence against their own mission obligations.

• Development firms

  Rangoon provides a launch architecture for development firms that need to design, review, and hand off agent capabilities while keeping client boundaries, approvals, and versioned delivery clear.

AI news and technical guides

• AI systems news and technical analysis

  AI ecosystem reporting and technical guides on agent systems, governance, infrastructure, and models.

• Agent systems insights

  Agent runtimes, handoffs, tools, and the boundaries that make them reviewable.

• Governance insights

  Policy, identity, authorization, approvals, and evidence for consequential work.

• Infrastructure insights

  Model serving, sandboxes, tool services, and deployment boundaries.

• Models insights

  Inference choices, local serving, compatibility, and operational tradeoffs.

• Editorial policy

  Rangoon Insights editorial policy for sourcing, dates, corrections, affiliation, and AI-assisted illustration.

• LNSAT: execution authority for AI agents

  How Rangoon separates capability management from LNSAT execution authority, exact action packets, approvals, receipts, and state reconciliation.

• Claude Sonnet 5.5: model upgrades require release controls

  Anthropic released Claude Sonnet 5.5 on September 28. Its arrival is a reminder to treat model upgrades as measurable, reversible application changes.

• Docker Cloud Sandboxes: moving agent workloads to managed compute

  Docker's Cloud Sandboxes extend its microVM model to managed compute. Teams now need to test portability, spend controls, and workflow handoffs.

• Canada and LawZero: planned investments and sovereignty claims

  Canada and Germany announced planned investments in LawZero. The announcement makes it useful to separate local capacity, research control, and data handling.

• GSA OneGov: consumption-based AI procurement

  GSA announced a new consumption-based OpenAI agreement for federal agencies. The practical challenge is governing usage, data handling, and evaluation.

• Gemini 3.8 Flash: model migration requires API controls

  Google released Gemini 3.8 Flash on September 2. The GA milestone puts version choice, lifecycle tracking, and workload testing back in focus.

• MCP's stateless core changes tool-service deployment

  MCP's July specification moved its core to stateless operation. The change makes ordinary web infrastructure a better fit for tool-service migration.

• DOE Genesis Mission: partner commitments and delivered capacity

  DOE reported more than $800 million in Genesis Mission partner commitments. The important distinction is between pledged capacity and delivered research capability.

• OpenShell and NemoClaw separate agent and runtime controls

  NVIDIA's OpenShell and NemoClaw split the agent harness from its runtime controls. That separation gives security reviews clearer seams to inspect.

• A2A v1.0 defines a stable agent handoff protocol

  A2A v1.0 stabilized an open protocol for agent-to-agent work. The milestone makes interfaces and version migration more concrete for builders.

• NIST's 2025 AISI-to-CAISI change and the current CAISSI mission

  Review the 2025 AISI-to-CAISI transition, NIST’s current CAISSI mission, and what published evaluation methods mean for technical and procurement teams.

• Docker Model Runner and the local inference boundary

  A practical guide to local model serving with Docker Model Runner, resource checks, API exposure, isolation, and evaluation discipline in practice.

• MCP tools: useful transport, separate authority

  Understand MCP hosts, clients, servers, tools, consent, and Rangoon’s exact-action authorization boundary before connecting agent systems safely.

• Ollama and vLLM: selecting an inference deployment

  Evaluate Ollama and vLLM by workload, hardware, API behavior, observability, and governance instead of treating local inference as one mode.

• OPA, Cedar, identity, and the exact action

  Compare policy engines and identity providers with execution authority, approvals, receipts, and the narrow scope of one consequential action.

Project information

• Community

  Rangoon invites future contributors to shape portable, governed agent-management systems through source review, practical extension work, and clear operational evidence.

• About Rangoon

  Rangoon is building toward an open, portable control plane for teams that want capable agent systems without treating configuration as unlimited authority.

• Website privacy

  How the Rangoon public website handles local interactions and requests.

• Website and preview terms

  Scope, development status, and rights information for the Rangoon website and design previews.

• Meet the Rangoon identity

  The Rangoon mascot, folded circuit symbol, and brand assets.

Discovery resources