Brenych Studio Agent Interface

Public MVP · September 3, 2026

Public MVP · Developer Tools · WebMCP

Brenych Agent Interface

Agent-native developer interface for structured, evidence-backed collaboration between AI agents and software projects.

  • Bootstrapped
  • Active development
  • Built by Brenych Studio
The Brenych Agent Interface after evaluation. A left column holds manually entered requirements — Electron, MCP and Supabase. The centre shows a dark core evidence graph with ranked project cards. A right column reports 100 percent evidence coverage and a requirement matrix with three of three requirements matched.
Fig. 01 The live interface after an evaluation run. Three manually entered requirements are matched against public evidence records, producing a ranked result set and a deterministic evidence-coverage readout. Scroll capture sideways for full detail

01 — What it is

A public WebMCP developer tool for agent-readable project evidence.

Brenych Agent Interface gives AI agents structured access to professional project context, capabilities and evidence. Instead of relying on unstructured portfolio browsing, an agent can discover the available capabilities, inspect evidence-backed project information, focus a relevant project and produce a structured collaboration brief.

People and agents work against the same public evidence records in the same live interface. Every claim in a result resolves to a record that can be inspected, and the interface states its own limits rather than implying capabilities it does not have.

02 — How it works

From project context to agent-readable interface.

Professional context is committed as structured evidence, exposed through a WebMCP tool surface, and consumed by an agent.

  1. 01

    Software project / professional context

    Real projects, delivered work and working history.

  2. 02

    Structured evidence

    Context committed as inspectable records with an explicit public boundary.

  3. 03

    Brenych Agent Interface

    The live surface where evidence is evaluated, ranked and focused.

  4. 04

    WebMCP tools

    Capabilities exposed to an agent as callable, discoverable tools.

  5. 05

    AI agent

    Discovers capabilities, inspects evidence and drafts a collaboration brief.

03 — Current MVP

What is live today.

The following is present in the public MVP and can be verified in the live application and its public source.

  • 01Structured professional profile
  • 02Agent-readable capabilities
  • 03Evidence-backed project context
  • 04Project discovery and focus
  • 05WebMCP tool interface
  • 06Structured collaboration brief generation
  • 07Public live application
  • 08Public source repository

04 — Product proof

The working interface, end to end.

Captures from the public MVP: the evidence field before evaluation, and the structured brief the interface produces once requirements are matched.

The interface before evaluation. The core evidence graph is marked as an unevaluated evidence field, with project cards for a creator operations workspace, a desktop agent interface, a desktop publishing workspace and further extended evidence records.
Fig. 02 The evidence field before evaluation. Project records load as an unevaluated core evidence graph — four evidence objects, two extended signals and one latent record — with manual requirement entry available alongside. Scroll capture sideways for full detail
The generated project brief. A working document lists project type, requirements, context, timeline, budget and the relevant evidence projects, under a status bar reading page-local draft only, no send, no CRM and no network write.
Fig. 03 The structured collaboration brief. Requirements, project type and the matched evidence projects are assembled into a working document. The interface declares its own limits in the header: page-local draft only, no send, no CRM, no network write. Scroll capture sideways for full detail

05 — Next stage

From public MVP to production.

The current public MVP establishes the agent-interface and WebMCP interaction model. The next stage focuses on production infrastructure, persistent workflows, observability and broader agent evaluation.

Planned — not yet implemented

AWS exploration

  • Serverless application infrastructure
  • API Gateway-based production interfaces
  • Persistent storage and workflow state
  • Operational observability
  • Amazon Bedrock-based agent and foundation-model evaluation

Brenych Agent Interface does not currently run on AWS. The public MVP is a static web application, and every item above describes intended next-stage work rather than shipped infrastructure.

06 — Verification

Public verification.

Public MVP
Launched September 3, 2026
Source
Public GitHub repository
Demo
Public video walkthrough
Development
Bootstrapped / founder-led
Status
Active development
OpenAI WebMCP Challenge
Submitted September 3, 2026