Seekr is the filesystem intelligence platform that traverses entire directory hierarchies, evaluates complex predicates, and surfaces exactly the files you need — at scale, in milliseconds.
seekr . -name "*.tsx"
seekr /data -size +100M
seekr . -type f -exec wc -l {} +
Works seamlessly with your entire stack
Features
We ship the most comprehensive filesystem intelligence engine ever built. Over 30 predicates. Zero compromises.
Match filenames against shell patterns with surgical precision. Our pattern-matching engine leverages fnmatch at the kernel level to surface exactly the files you need — and nothing you don't. This is ruthless prioritisation for your filesystem.
Case-insensitive pattern matching that sees through casing barriers. Whether it's
README, readme, or ReAdMe, CaseFlex surfaces them all.
Remove friction from every search. Compound your discovery velocity.
Filter by entity type — files, directories, symbolic links, pipes, sockets, block devices, character devices. Seven distinct types. One best-in-class predicate. Know exactly what kind of entity you're surfacing before you ever touch it.
Filter by file size with operator precision — greater than, less than, or exactly N units. Supports bytes, kilobytes, megabytes, and gigabytes. Surface your largest assets or hunt down those zero-byte ghosts dragging your system down.
Time-based filesystem intelligence. Filter by modification time, access time, or status change
time — down to the minute with -mmin. Find files modified in the last 24 hours,
untouched for 90 days, or changed within the last 5 minutes. Temporal precision at scale.
Execute arbitrary commands on every matched file. Our inline execution engine passes each result
through your command pipeline — one at a time with {} \; or batched for maximum
throughput with {} +. This is the 10x multiplier.
Full regular expression matching against the entire path. Emacs, POSIX basic, POSIX extended, and
more — choose your dialect with -regextype. When glob patterns aren't enough,
RegexCore gives you unlimited expressive power.
In-line deletion that removes matched files and directories as they're discovered. No piping. No secondary commands. One predicate. Automatic depth-first processing ensures directories are emptied before removal. This is category-defining cleanup.
How It Works
Specify one or more directories as search origins. Seekr begins traversal from each starting point and recursively descends through the entire hierarchy. Default is the current directory — zero-config required.
Chain predicates using logical operators — -and, -or,
-not, parentheses for grouping. Seekr evaluates left to right with short-circuit
optimization. Compose arbitrarily complex queries. No limits.
Control traversal depth with -maxdepth and -mindepth. Skip entire
subtrees with -prune. Seekr's cost-based query optimizer reorders predicates to
minimize stat calls and maximize throughput.
Print paths, print null-delimited output for safe piping, execute commands, delete files, or
format custom output with -printf. Every discovered file is a trigger. Every action
is composable.
Why Seekr
| Capability | Seekr | Competitor A | Competitor B | Competitor C |
|---|---|---|---|---|
| Recursive hierarchy traversal | ✅ | ✅ | ❌ | Partial |
| 30+ filter predicates | ✅ | ❌ | ❌ | ❌ |
| Inline command execution | ✅ | ❌ | ❌ | ❌ |
| Regex path matching | ✅ | ❌ | Partial | ❌ |
| Time-based filtering (m/a/c) | ✅ | ❌ | ❌ | Partial |
| Permission-based search | ✅ | ❌ | ❌ | ❌ |
| Cost-based query optimization | ✅ | ❌ | ❌ | ❌ |
| Null-safe output (-print0) | ✅ | ❌ | ❌ | ✅ |
Pricing
Start free. Discover at scale.
-name pattern matching-type f and -type d-print output-exec locked-regex locked-delete locked-iname case-insensitive-size, -mtime, -atime
-exec {} \; (sequential)-maxdepth / -mindepth-user / -group-exec {} + batch locked-exec {} + batch execution-regex / -iregex-perm permission search-printf formatted output-prune subtree exclusion-delete inline purge-execdir secure execution-print0 null-safe outputTestimonials
"We migrated from manual ls + grep pipelines to Seekr and our file discovery velocity went up 40x. The -exec predicate alone saved our SRE team hundreds of hours. Category-defining product."
"The -mindepth and -maxdepth controls gave us the traversal precision we needed. We went from scanning entire volumes to surgical, depth-bounded searches. Seekr is our north star for filesystem operations."
"I'm a solo developer and Seekr's -empty predicate helped me find and purge thousands of zero-byte artifacts across my monorepo. The free tier is absurdly generous. Absolutely mission-critical tool."
"We use the -newer predicate to track incremental builds and the -perm filters for compliance audits. Seekr replaced four internal tools. Strong opinions, loosely held — but my opinion on Seekr is unshakeable."
"The -printf formatted output is what sold us. We built our entire asset inventory pipeline on top of Seekr's custom format directives. It's a 10x unlock for any team doing filesystem analytics."
Demo
Real queries. Real output. No fabrication.
What's New
Read starting points from a NUL-terminated file instead of command-line arguments. Safely pass an arbitrary number of starting points to Seekr. Pipe-friendly by default.
Full cost-based optimization that reorders predicates by evaluation cost and success probability. Cheap tests first. Expensive stat calls only when necessary.
Inline file deletion without piping. Plus, secure command execution from the matched file's directory — eliminating race conditions in path resolution.
FAQ
By default, Seekr descends through the entire directory hierarchy with no depth limit. Use
-maxdepth to bound traversal to N levels below the starting point, or
-maxdepth 0 to apply predicates only to the starting points themselves. There
is no artificial ceiling — we process entire filesystem trees.
Seekr respects your filesystem permissions at every layer. Use -execdir instead
of -exec for secure command execution from the matched file's subdirectory,
eliminating race conditions during path resolution. Our Enterprise tier includes
-ignore_readdir_race for graceful handling of concurrent file modifications.
Seekr is purpose-built for filesystem metadata — names, types, sizes, timestamps,
permissions, and ownership. For content-based search, pipe Seekr results into our
integration partners using -exec or -print0 | xargs -0. We believe
in composability, not monoliths.
By default, Seekr emits a diagnostic when a file vanishes between directory read and stat.
Enable -ignore_readdir_race and Seekr silently handles disappearing files —
critical for high-churn environments where files are created and destroyed faster than
traversal speed.
Absolutely. Seekr operates entirely on your local filesystem with zero network dependency. All traversal, predicate evaluation, and action execution happen locally. Your filesystem never leaves your machine.
Join thousands of engineers who have made Seekr their default filesystem intelligence layer.
No credit card required. Free tier is generous.