Skip to content

Altimate Code Chat

What is Altimate Code?

Altimate Code is the open-source data engineering harness with 100+ deterministic tools for building, validating, optimizing, and shipping data products. It is now integrated into the Datamates extension, giving you AI-powered data engineering chat directly in your IDE.

Getting Started

Open Altimate Code Chat

  1. Press Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows/Linux) to open the command palette
  2. Type Datamates
  3. Select Datamates: Open Altimate Code Chat

This opens the Altimate Code chat panel where you can interact with agents and use all 100+ data engineering tools alongside your Datamates integrations.

Features

Agent Modes

Altimate Code provides three agent modes to match your workflow:

Mode Access Level Use Case
Builder Full read/write Scaffolding dbt projects, writing models, generating tests and docs
Analyst Read-only Exploring schemas, running queries, analyzing lineage
Plan Minimal access Planning changes, reviewing impact before execution

100+ Data Engineering Tools

  • SQL Tools — Validation without execution, query optimization, anti-pattern detection, dialect translation, PII scanning
  • dbt Tools — Model generation, test generation, documentation generation, project scaffolding, troubleshooting
  • Lineage Tools — Column-level lineage, impact analysis, downstream dependency tracking
  • Schema Tools — Schema exploration, table/column discovery, metadata indexing
  • FinOps Tools — Cost analysis, warehouse spend reports, optimization recommendations
  • Warehouse Tools — Direct query execution, result preview, connection management

Works with Datamates Integrations

Altimate Code Chat has full access to your configured Datamates integrations — Snowflake, BigQuery, Databricks, PostgreSQL, dbt, Airflow, Dagster, Jira, Linear, GitHub, and more. The chat agent can query your warehouse, inspect your dbt project, check your Jira tickets, and review your GitHub PRs — all from a single conversation.

Benchmarks

Altimate Code is #1 on ADE-Bench — the industry benchmark for AI data engineering agents, created by Benn Stancil (founder of Mode) in collaboration with dbt Labs. It evaluates agents on real-world analytics and data engineering tasks using actual dbt projects and databases.

The harness — not the model — is the differentiator. Despite using Sonnet 4.6 (not the most expensive model), Altimate Code outperforms agents running on more capable models, demonstrating that purpose-built tooling and deterministic operations outperform raw model capability alone.

ADE-Bench (DuckDB Local)

Tool Model Score Pass Rate
Altimate Code Sonnet 4.6 32/43 74.4%
Cortex Code CLI Opus 4.6 28/43 65%
dbt Labs Sonnet 4.5 ~25/43 59%
Claude Code (baseline) Sonnet 4.6 ~17/43 40%

Other Benchmarks

Benchmark Result
SQL Anti-Pattern Detection 100% accuracy across 1,077 queries, 19 categories. Zero false positives.
Column-Level Lineage 100% edge match across 500 queries with complex joins, CTEs, and subqueries.
Snowflake Query Optimization (TPC-H) 16.8% average execution speedup (3.6x vs baseline).

Full benchmark details →

LLM Access

Two options for powering the AI chat:

  • BYOK (Bring Your Own Key) — Free and unlimited. Use any of 35+ supported providers (Anthropic, OpenAI, AWS Bedrock, Azure OpenAI, Google, Ollama, and more)
  • Altimate LLM Gateway — Managed LLM access with dynamic routing across Sonnet 4.6, Opus 4.6, GPT-5.4, GPT-5.3, and more. 10M tokens free to get started — no API keys to manage

Standalone Usage

Altimate Code can also be used outside the IDE as a standalone CLI/TUI:

Interface Description
TUI Interactive terminal UI — altimate
CLI Command-line for scripting — altimate run
Web UI Browser-based interface — altimate web
CI/CD Headless mode for pipelines — altimate check
GitHub/GitLab Automated PR review and issue triage

Install standalone:

npm install -g altimate-code

Full Documentation