Comparison

A flattener gives you a file. IngestMD gives you the right context.

Tools like Repomix and GitIngest are genuinely good at one thing: turning a repo into a single file, fast. IngestMD does that too, then solves the next problem, handing AI exactly what a job needs without dumping everything and burning your token budget.

If all you want is the whole repo in one text file to paste into a chat, a flattener is the right tool, and IngestMD will do it in one click too. The difference shows up on real work: a large codebase where the model needs one subsystem, a review that should see the tests and dependents, a client folder that cannot leave your machine, a pile of documents, or a Salesforce org where the folders do not match the features.

Side by side

By capability

Capability A flat repo dump IngestMD
What you get The whole tree flattened into one file, fast. The same fast bundle, plus task-scoped and area-scoped context when you want the right slice, not everything.
Choosing what goes in You hand-tune include and exclude rules until the file is small enough. Describe a task and it ranks, dependency-expands, and packs the files to a token budget, with a reason for each.
Understanding structure A file list. The model has to infer how the parts relate. Area discovery clusters the repo into real subsystems by imports and shared terms, one titled bundle each.
Safety Secret and binary handling varies by tool and config. Secret redaction, binary skipping, and a NUL-byte guard are on by default and never paywalled.
Beyond code Source files only. PDF, Word, and Excel into clean Markdown, plus Confluence retrieval, redacted the same way.
Reuse Re-run the command each time. Curate a Collection, save it by name, and reuse or serve it to AI clients over MCP.
Salesforce A folder of metadata, ungrouped. Groups Apex, LWC, and objects into the features they form, following @salesforce/apex calls and SOQL.
Where it runs CLI or a hosted web page. Native desktop app, CLI, and MCP server, the same engine across all three.
Privacy Local or hosted depending on the tool. Local by default, nothing uploaded. Network egress is opt-in: your own local LLM, optional cloud models with your key, or Confluence.

Comparison is to the general "flatten a repo" approach. Specific tools differ, and several have their own strengths, check their docs for current features.

Honest take

When to use which

Reach for a flattener

  • You want the whole repo in one file, right now.
  • It is a small project that fits the context window whole.
  • You just need a quick paste into a chat.
  • You are grabbing a public repo by URL.

Reach for IngestMD

  • The repo is too big to dump, and you need the right slice.
  • You want AI to review changes with their tests and dependents.
  • The code or documents are confidential and cannot be uploaded.
  • You are working in Salesforce metadata.
  • You want to curate, save, and reuse context, or serve it over MCP.

Try it on a repo that is too big to dump

Free to download, with a 3-day full Pro trial. No account, no card, nothing leaves your machine.

Download IngestMD See every feature