Atlas Convention Drift compares the coding standards you wrote down against what your code actually does. When a recently changed file breaks a rule you declared, such as one concern per file or no em-dashes in client-facing copy, Atlas surfaces the violation in a sidebar panel while it is fresh. The scan contains no LLM calls at all, so it costs nothing to run. The engine ships in Atlas today behind the paid tier, and Atlas Pro is not yet on sale.
Standards you declared, quietly contradicted
You write down how the code should be built, and then it drifts 1 file at a time, until review or never.
Convention Drift treats each declared convention as a claim and your recently changed code as reality. When a new change breaks a rule, that violation surfaces while you can still fix it cheaply, instead of accumulating unseen until someone reads the whole diff.
Zero tokens, by design
The entire scan runs with 0 LLM calls, which is unusual enough in 2026 to be worth stating plainly.
Each convention is a deterministic checker: regex rules such as no em-dashes and no emoji, count rules such as file-too-long, and structural heuristics such as one concern per file. Because none of it calls a model, Atlas can run the check continuously without metering you. Deriving new conventions from your project guide's prose would need a model, and that part is designed but not yet built.
New drift, not old debt
Flagging an entire codebase at once is noise, so the first run looks back only 14 days by default.
After that, Convention Drift scopes its checks to files changed since the last watermark, so the nudges are about fresh drift you just introduced rather than a backlog you already know about. Each violation carries a stable identity, so acknowledging one keeps it quiet while a genuine reintroduction fires again.
Availability
Convention Drift is 1 of the 8 record engines in Atlas, and it sits behind the paid tier.
The engine is built and running in the product, gated to Atlas Pro. Pro has no checkout yet, so it is not something you can buy today. Shipped, the automated work log, is the one record engine available free right now.
Frequently asked questions
- Is Atlas convention drift detection available now?
- The engine is built and running in Atlas, but it sits behind the paid tier and Atlas Pro is not yet on sale, so you cannot purchase access today.
- Where do the conventions come from?
- Built-in checkers for common rules, plus an optional conventions file in your project. Deriving conventions from your project guide's prose is designed but not yet built.
- Does it check my whole codebase?
- No. The first run looks back 14 days by default, and after that it scopes checks to files changed since the last pass, so it flags new drift rather than legacy debt.
- Does it use an LLM?
- No. The scan contains zero LLM calls and is fully deterministic, which is why it can run continuously without costing you anything.
- Will it edit my code to fix violations?
- No. It surfaces violations with file and line in a sidebar panel. You decide what to change.
Try Atlas in your terminal
The terminal-native AI coding agent. Free core, single binary.
Install AtlasRelated guides
Atlas for Bash in 2026
Atlas is a terminal-native AI coding agent for Bash in 2026, with shellcheck-aware suggestions. Harden a script or fix shellcheck warnings and review every diff.
Atlas with Gemini Flash Latest: The Rolling Alias Explained (2026)
gemini-flash-latest in Atlas: a rolling alias, not a pinned checkpoint. $0.3 per Mtok input, $2.5 per Mtok output, 1,048,576 token context, and no reproducibility.
Atlas with Qwen3 32B: The Largest Dense Qwen3 in 2026
Qwen3 32B in Atlas: 16,384 max output tokens, hybrid thinking, 128K tokens (131,072) of context, $0.70 per Mtok input and $2.80 per Mtok output in 2026.
Atlas with Gemini 2.5 Pro: The Cheap 1M Context Default in 2026
Gemini 2.5 Pro in Atlas: a stable GA id with a 1,048,576 token window at $1.25 per Mtok input, 37 percent cheaper to read with than Gemini 3 Pro.
Atlas with Amazon Nova Micro in 2026: The Cheapest Model on Bedrock
Amazon Nova Micro costs $0.035 per Mtok input, the lowest price in the Bedrock catalog, with a 128K token context. Use it as Atlas's small_model, never as the build loop.
Atlas vs Magic.dev: Terminal AI Coding Agents in 2026
Compare Atlas, the terminal-native AI coding agent with permission-gated tools and diff review, against Magic.dev's research claims of 100 million token context models in 2026.
Onboard to an Unfamiliar Codebase with Atlas in 2026
How to onboard to an unfamiliar codebase with Atlas in 2026: use codebase_search, glob, read, lsp, task, and todowrite to build a mental model fast.
Atlas with DeepSeek Chat: 384,000 Token Output at $0.28 per Mtok in 2026
Run Atlas on DeepSeek Chat in 2026. DeepSeek's non-reasoning endpoint gives 1M tokens (1,000,000) of context at $0.14 per Mtok input, $0.28 per Mtok output.