Skip to content

> cat /etc/evolution.log

What Maguyva runs on

Built for speed, debuggability, and continuous improvement. The stack is boring where it should be boring and explicit where generic abstractions made the agent workflow worse.

The current production architecture grew out of one constraint: make agent workflows easier to debug under pressure, not more abstract.

PostgreSQL + Lance Quantized vector search Code language engine MCP everywhere 229+ languages

Production Layers

> lsmod | grep production

These are the core layers that stayed after multiple iterations. Every one of them earned its place by making debugging easier or by removing friction from agent execution.

Database

PostgreSQL + Lance

Tenant-partitioned storage and indexes. Your code stays isolated, with no cross-user index mixing.

Embeddings

Voyage AI voyage-code-3

Premium code embeddings for your agents, supplied by Voyage Code 3

Vector Search

Binary Quantized HNSW

Embeddings compressed to 2048-bit signatures and indexed with HNSW. 32x storage reduction, sub-millisecond search, zero extra infrastructure.

Language Engine

Code Language Engine

A heavily modified Tree-sitter stack with custom grammars, handler heuristics, and 229-language extraction coverage

Protocol

MCP Protocol

Remote MCP server for Claude Code, Claude Desktop, Cursor, VS Code, Windsurf, and other MCP-compatible clients. 11 tools exposed via FastMCP.

Cache

Lance columnar + R2

Local-first, cloud-synced embeddings

Architecture Timeline

The stack after the dead ends

Maguyva did not arrive at this shape all at once. Each layer survived because it made the system faster to reason about, cheaper to operate, or easier to debug under pressure.

The orchestration experiments started earlier. This timeline starts in September 2025, when the Maguyva product itself snapped into focus.

1
2025-Q3 foundation

Focused Maguyva build starts → Voyage voyage-code-3

The product direction hardens and the rapid iteration loop begins

2
2025-Q3 storage

Binary vector retrieval layer

Quantized search with compact indexes, no separate vector stack

3
2025-Q4 ast-engine

Code language engine comes online

Custom queries, parser fallbacks, and handler-based heuristics across 229 languages

4
2025-Q4 search

Multi-modal fusion search

Ranked retrieval across semantic, text, AST, and graph indexes in a single query

5
2026-Q1 caching

Lance columnar + R2 sync

Local-first, 10x faster cold starts

6
2026-Q1 quality

229-language validation framework

Automated extraction testing across every supported language

NOW launch

Public launch at maguyva.ai

7 months from focused build to public availability

Customer Zero

> ps aux | grep maguyva

Maguyva is built by agents that run on Maguyva.

Our orchestration system coordinates specialist AI agents that use the same MCP tools, search indexes, and production infrastructure you get. There is no internal version. The team page lists “1x Maguyva ($49) — we pay ourselves.” That is not a joke. It is the invoice.

// customer zero ships to production daily

Design Principles

> cat /etc/principles.d/*.conf

These are the constraints that stayed useful in production. They are less about ideology and more about keeping the system understandable while agents are making changes at speed.

!

No Silent Fallbacks

Fail fast, fix fast. Every error surfaces immediately. No graceful degradation hiding bugs in production.

// panic > silent corruption

λ

RPC-First

Database-side logic for performance. PostgreSQL functions execute close to data, eliminating round trips.

// PostgreSQL functions, 0 ORMs

CQRS

Separate write and read paths. Indexing pipeline writes, MCP tools read. Different optimization strategies.

// writes: batched | reads: cached

#

Deterministic UUIDs

Cache coherence without database lookups. Same input = same ID. Idempotent operations by design.

// uuid5(namespace, content_hash)

The best architecture is the one you can debug at 3am.
(We've tested this.)

// last updated: 2026-03

Try it on your repo →

MAGUYVA.NFO