AI that understands enterprise systems.
Recover business logic from legacy ETL, SQL, and enterprise code.

Companies have all of this. None of it is usable by AI yet.

Not slides. Real runs against real repos.

A compiler for enterprise knowledge — not another RAG index.

Run cold against real repos, not staged demos.

Every deck below reproduces its claims live — real terminal output, real MCP query/answer transcripts, and the honest gaps found along the way, never papered over.

We grade our own agent. The failing numbers stay on the page.

A 101-scenario suite across seven categories, graded by six deterministic evaluators — no LLM judge. It is run against the real compiled ledger, and it is the reason the numbers below moved: the harness found that 94% of what we were compiling wasn't our own code (a Python virtualenv and two scanner-output directories nobody had excluded).

That last tile is the point. A benchmark you publish only when it flatters you is marketing, not measurement — so the regression ships alongside the gains, and the suite still reports Status: FAIL against its own gates. The full report keeps every superseded number visible as a correction rather than editing it out.

Built in the open. Improved in the open.

Presentations

• Full-Stack Test Run

  An autonomous 22-act pass over the distributed storage stack, the compiled-knowledge query engine, and the MCP protocol — coordinator fencing, gateway failover, RRF fusion, REASON citation checks, the vector arm on a real embedding model, and the write-vs-read-only-gateway safety assertion. Three partitioned-store bugs it surfaced, fixed and re-verified in the same run.

  New · full-stack verification · RFC 0111–0126 / 0013 / 0115
• Distributed Storage Under Fire

  Two autonomous end-to-end test runs of EKOS's object-storage-backed, horizontally-distributed knowledge ledger — fault injection at the coordinator, the compile worker, and the query gateway, against a real S3 endpoint and a 95-partition workspace. Eight defects found and fixed with regression tests; every metric traced to a log file.

  New · distributed storage · RFC 0111 / 0113
• EKOS Compiles Itself

  The full pipeline, storage footprint, three documentation-generation layouts, the whole MCP tool surface, and real local-LLM token usage — every number measured against EKOS's own real ~50-crate repository, not estimated. Includes 3 real bugs found and fixed live during the run. See also the full capability reference — every RFC and command above, plus live-generated evidence from this exact repository.

  New · self-hosted
• The AI-Native Enterprise Knowledge Compiler

  The startup pitch: why enterprise AI needs a compiler, not another RAG index — proven cold against real repos, and audited live by Claude Code using EKOS's own MCP server on itself.

  Start here
• Reproduce Legacy ETL Logic, Verified

  Pentaho jobs, SQL, and stored procedures compiled into one evidence-backed Transformation IR — explain what a legacy pipeline does, and prove a migration didn't change its meaning.

  RFC 0027 · RFC 0028 · RFC 0029
• Compile GitHub Repository → MCP Server

  Issues, PRs, files, and their real relationships compiled into an append-only, evidence-backed ledger any MCP-speaking agent can query.

  RFC 0013 · RFC 0020
• GitHub, Live, End to End

  First live run — 1,600 real issues/PRs, a real repo, not a mock. Two known gaps fixed before the run; a third (96% of items collapsing into one identity) found only at real scale and fixed the same session. Real docs, code, and GitHub items surfaced by one query — plus the honest residual limitation, not hidden.

  RFC 0020 · RFC 0062
• Pentaho + SQL Recovery, Tested Cold

  EKOS's Pentaho and SQL recovery analyzers run against two real GitHub ETL projects — terminal output, MCP query/answer transcripts, and the honest gaps a data engineer would want to see before trusting it.

  RFC 0025 · RFC 0027
• Two Real Recovery Gaps, Filed and Closed

  The honest gaps the previous deck surfaced became GitHub issues #2 and #3 — this deck reruns the exact repro shapes against the fixed pipeline, live, with the before/after coverage numbers.

  RFC 0031 · issues #2, #3
• Claude Code + EKOS

  How Claude Code searches and analyzes a codebase through EKOS's MCP server instead of raw grep/Read — a measured, honest with-vs-without comparison, the real /context token footprint, real 24-hour usage logs, and gaps stated plainly.

  RFC 0013
• Generated Documentation

  Markdown, self-contained HTML, and Mermaid diagrams rendered straight from the compiled ledger — including a real cloned Pentaho ETL repo compiled cold, join and Unmapped steps included. Zero LLM by default; an opt-in prose tier verified against a real local LLM. Every example is real output, unedited.

  RFC 0035
• Cross-Project Generated Documentation

  The same ekos docs generate pointed at EKOS's own ~50-crate workspace, a real Databricks Asset Bundle repo, and a real Azure Data Factory pipeline repo in one session — real counts, real links, and an honest gap called out where no ADF-specific connector exists yet.

  RFC 0035 · RFC 0037 · RFC 0040 · RFC 0041
• ClickHouse: Compiled Metadata + Live NL-to-SQL

  The one explicit, audited exception to "AI never touches raw enterprise systems directly" — an LLM-built, SELECT-only, redacted, ledger-audited live query, verified against a real ClickHouse container, honest failures included.

  RFC 0056
• EKOS Cold Against Plausible's ClickHouse Layer

  Pointed at a real, unmodified open-source repo it had never seen, EKOS compiled its git/CI/dependency knowledge cleanly and found its own gap: sqlparser's ClickHouseDialect can't parse ClickHouse's own CODEC(...) column clause. Documented anyway, gap named plainly.

  RFC 0031 · RFC 0056
• ClickHouse Extraction, After the Fix

  The same repo, re-analyzed after RFC 0057/0058 closed the parser gap: all 15 real ClickHouse tables now structurally recover with zero warnings — and re-analyzing surfaced a second, unrelated finding in a different pipeline stage, identity resolution over-merging 6 of those 15 tables into one. Reported the same way the first gap was.

  RFC 0057 · RFC 0058
• Vision & Token Utility

  Why the EKOS token's relevance is designed to grow as a consequence of platform adoption, not a promise of price — the phased ecosystem roadmap from contributor rewards to plugin, agent, and knowledge marketplaces.

  VISION.md

The people building EKOS.

Alexey Banaev

Founder & Lead Developer

Alexey is a Data Engineer and Solution Architect with extensive experience designing data platforms, analytics solutions, and distributed systems. He founded EKOS to address a fundamental challenge in modern data environments: making complex, fragmented data ecosystems understandable, traceable, and usable.

As EKOS's founder and main developer, Alexey leads the product architecture and engineering, combining data engineering, software development, and AI-driven approaches to build a practical solution for modern data teams.

Omid Ahmadi

Core Team, Ecosystem & Growth · AI Researcher & Agentic Systems Engineer

Researcher focused on sustainable AI applications across the oil, gas, and petrochemical industries, with a focus on intelligent agents, autonomous systems, and reliable AI infrastructure for industrial environments.