Working systems
Products and developer tools with concrete capabilities, limitations, and implementation details.
Altru.dev · Code for Humanity
Building what’s next.
Independent software, developer tools, assurance architecture, applied research, and public technical progress.
BUILDVERIFYPUBLISHIMPROVE
Now
All Progress
Current development
From Permission to Proof: Frequency Agent Execution Assurance
Frequency Agent Execution Assurance separates permission from execution evidence: requested action, effective authority, observed side effects, drift, and durable receipts. MCP Lens compleme
Read latest update →Free Connectivity Is Not the Same as Trusted Connectivity
Frequency SafetyProtocol explores how legitimate free and shared connectivity can be used safely by separating network usefulness from network trust, protecting public-network traffic, minim
Read latest update →The Failure Didn’t Disappear. It Moved.
Recent work on agentic systems keeps exposing the same pattern: improving one layer rarely removes failure; it often relocates it. The practical problem is to preserve uncertainty, authority
Read latest update →Technical depth
Architecture and assurance work documented closely enough to understand what changed and why.
Claims stay bounded
Trust, privacy, verification, and release information are shown as structured facts rather than generic promises.
Projects
View All ProjectsReleased and active systems.
Concrete capabilities, current state, implementation details, and direct product or source links.
Prototype
Altru Browser
Independent adaptive browser and native web engine. Rust-first, cross-platform, privacy-minded, and evidence-gated.
App
Published
Living Architecture Nodes
Keep architecture memory beside the code: find missing or stale nodes, review drift, and export clean AI/dev handoff context.
Extension
Public Beta
Frequency Agent Execution Assurance
Govern what an AI agent is allowed to do, observe what it actually did, and preserve evidence of the resulting external state.
Service
Beta
Agent Replay
Standalone forensic reconstruction for AI-agent executions. Rebuild what happened, identify the first provable divergence, and preserve the evidence supporting that conclusion.
Service
Beta
MCP DriftGuard
Continuous integrity verification for MCP servers. Detect whether the security-relevant server running now still matches the server that was originally approved.
Service
Beta
BoundaryProof
Independent assurance for AI-agent execution boundaries. Compare intended, authorized, and observed state transitions to determine whether execution stayed within its permitted boundary.
Service
Beta
RenderDiff
Evidence-first representation-divergence analysis for text, documents, and HTML. Find material differences between what humans see and what software actually receives.
App
Beta
DDC PR Proof
Semantic proof-of-change for pull requests. Identify changes in behaviour, authority, failure semantics, assumptions, and guarantees that may not be obvious from the code diff.
Service