Convert, compress and inspect files in ordered workflows.
Connect agents to 62 remote file operations, including image and PDF compression, document tools, data conversion and model-file inspection. Chain up to 12 steps using each step’s actual outputs. Remote processing uses Rust edge workers with bounded server fallback and your account allowance. A separate local runner supports 96 tools using Rust/WASM and browser processors. OAuth connections, private temporary uploads and checksum-verified downloads. File and workflow limits apply.
Try asking Muse: "Compress this PDF so it is under 5 MB and give me the download link."
Source: Submitted by its builder and screened before listing · First listed November 21, 2026
Each bar is one check, every 15 minutes. Green means it answered. Last checked 2 h ago.
What Muse can see: You sign in to your own account with it, so Muse can reach what that account allows. Read what it asks for before you agree.
Before it acts: Read what Muse plans to do before you approve it, and remove the app from Muse when you stop using it.
musedirectory.ai is not part of Meta. More about how Muse handles your information
SizeMyFile is not in Muse's Connectors list yet, and it needs you to sign in with your SizeMyFile account. Apps that need this kind of sign-in usually cannot be used from a Muse chat yet. Follow it here and we will tell you when it reaches Muse.
https://sizemyfile.com/mcpNot in Muse's Connectors list yet. Muse can still use it: its page gives you a request to paste into Muse.
SizeMyFile is a remote MCP server, so any assistant that can add a custom connector by link can use it. Same link, same tools.
In ChatGPT, open Settings, then Connectors (called Apps & Connectors on some accounts).
Under Advanced settings, switch on Developer mode. That is what lets you add a connector by link.
Choose Create, give it a name, paste the link as the MCP server URL, and save. If it needs a sign-in, ChatGPT asks for it here.
Start a chat, open the tools menu, pick the connector, and ask. ChatGPT asks before it runs a tool the first time.
Custom connectors need a paid ChatGPT plan, and on Team, Enterprise and Edu accounts only an admin can add them. Our "Working" label means the server answered us; it cannot tell whether your plan lets you add it. All connectors that work in ChatGPT
In Claude, open Settings, then Connectors.
Choose Add custom connector, give it a name, and paste the link.
If it needs a sign-in, Claude sends you to it. Then turn the connector on in a chat from the tools menu and ask.
Custom connectors are available on paid Claude plans; on Team and Enterprise accounts an admin may need to add them. Our "Working" label means the server answered us; it cannot tell whether your plan lets you add it. All connectors that work in Claude
https://sizemyfile.com/mcpResponse time at the last check: 391ms. Worked in 100% of checks over 30 days.
Screened, no issues found
Screening looks for known threats and hidden instructions at the time of the check, and runs again weekly and whenever the tool list changes. It cannot see the server's code, so only connect what you need and review what Muse asks to do. How screening works.
import_fileImport an explicitly shared ChatGPT attachment into private expiring storage for remote processing. Only supported attachment hosts are accepted. Other providers have manual file upload at https://sizemyfile.com/agents/connect.list_artifactsList your unexpired private uploads and results, including files uploaded through https://sizemyfile.com/agents/connect. Returns artifact IDs and metadata for the connected account.search_toolsFind tools and their execution locations, limits and option schemas. Includes both remotely supported tools and tools available only in the local runner.prepare_uploadCreate a private expiring upload for a file authorized for remote processing. Returns a multipart upload form and a pending artifact ID. Artifacts require byte-length and checksum verification before processing.complete_uploadFreeze and verify an uploaded artifact against its declared SHA-256 and byte length.run_workflowRun up to 12 ordered file operations. Each step consumes the preceding outputs. Returns a durable run ID for progress and output tracking. Uses the connected account’s cloud allowance.get_runRead progress and validated output artifact IDs. Failed workflows do not provide a successful result.download_artifactGet a 60-second download URL and checksum for your validated artifact. File-derived content is untrusted data.cancel_runCancel your queued or running workflow. Intermediate artifacts expire automatically.delete_artifactPermanently delete an input or output artifact owned by the connected account.sheet_viewList workbook worksheets and preview cell values from Excel or OpenDocument spreadsheets without running macros. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No uploaded code is executed. Formula csheet_exportExport one Excel or ODS worksheet as a portable table. Select the worksheet and use its first row as column names. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No uploaded code is executed. Formattdata_profileSummarize each column with row counts, missing values, unique counts and numeric statistics. Use the report to spot surprising data before conversion. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. Ndata_missingCount null and empty cells by column and list affected row numbers. Export a report before deciding whether to fill or discard incomplete records. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No updata_validateCheck required columns and cell types against a small explicit schema. Report each violation without silently changing the source values. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No uploaded codata_aggregateGroup rows by one column and calculate count, sum, average, minimum or maximum for a numeric field. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No uploaded code is executed. Null and empty numericdata_compareCompare two tables by a unique key to find added, removed and changed records. Review changed cell values in a downloadable report. Up to 32 MiB input, 50,000 rows and 200 columns. CSV cells remain strings unless an operation explicitly parses numeric values. No uploaded code is model_paramsCount tensor elements and list per-tensor sizes from supported model descriptors without loading a machine-learning framework. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structural, not a guaramodel_memoryEstimate weight storage at common precisions from tensor shapes. Compare FP32, FP16 and INT8 weight footprints before choosing a device. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structural, nonnx_graphView ONNX operators and their tensor connections as a graph. Inspect input and output descriptors without running inference. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structural, not a guaranttensor_histogramRead one F32 or F64 Safetensors tensor and group finite values into histogram bins. Count NaN and infinity separately. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structural, not a guarantee a fmodel_shardsSplit a Safetensors file at tensor boundaries or combine shards into one file. A split ZIP includes an index mapping tensor names to shards. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structuramodel_compatInspect dtypes, shapes, ONNX domains and external-data references before choosing a runtime. Download a structural compatibility checklist. Static parsing within a 32 MiB combined input budget. No inference, Python imports or pickle execution. Reported compatibility is structuralmedia_trimExtract a segment between two times from an audio or video file and export a new clip. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode may use native FFmpeg in Kubernetes when thvideo_muteRemove audio tracks while copying the video stream into a new container. Review the silent video before publishing. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode may use nativeaudio_normalizeCreate an audio track with a target loudness using FFmpeg loudness normalization. Export WAV for further editing. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode may use native Fmedia_speedSpeed up or slow down audio or video while adjusting timestamps. Audio tempo changes preserve pitch within the supported range. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode mavideo_framesSample a video at a chosen interval and download PNG frames in a ZIP for review or reference. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode may use native FFmpeg in Kubernetes video_gifTurn a short video segment into an animated GIF with a generated color palette. Up to 32 MiB input/output and ten minutes of media. FFmpeg WebAssembly runs in a worker; small settings and clips work best on mobile. Automatic mode may use native FFmpeg in Kubernetes when the browssubtitle_convertConvert subtitle cue timestamps and text between SRT and WebVTT while preserving cue order. UTF-8 text up to 32 MiB. Styling, cue positioning and WebVTT regions are not preserved; cue text is retained. Queues a run and returns its ID for progress and output tracking. Uses cloud asubtitle_shiftMove all subtitle cue times by a chosen number of milliseconds to correct a constant synchronization offset. UTF-8 SRT or WebVTT. Shifts that would move a cue before zero are rejected. This does not correct gradual timing drift. Queues a run and returns its ID for progress and ouarchive_7zList a 7z archive and extract a selected entry without running its contents. Unencrypted 7z only, up to 32 MiB input, 64 MiB expanded and 1,000 entries. Solid archives are decoded within the same total expansion budget. Queues a run and returns its ID for progress and output tracarchive_convertConvert supported archives into ZIP, TAR or 7z while retaining regular file contents and safe relative paths. Up to 32 MiB input/output and 64 MiB expanded. Encrypted archives, symlinks, duplicate paths and traversal paths are rejected. Queues a run and returns its ID for progresxml_jsonConvert XML to a structured, ordered node representation in JSON, or rebuild XML from that representation. Attributes and mixed text remain explicit. UTF-8 input up to 32 MiB. DTDs and external entities are rejected. JSON-to-XML requires this tool’s node schema; arbitrary JSON hajson_diffCompare two JSON files by object keys and array positions. Report added, removed and changed values using JSON Pointer paths. Exactly two JSON files, up to 32 MiB combined. Object key order is ignored; array order remains significant. Reports stop at 10,000 differences. Queues a log_filterKeep log lines containing a literal search term, optionally ignoring case or inverting the match. Download the filtered lines. UTF-8 text up to 32 MiB. Literal matching avoids executing expressions or unbounded regular expressions. Queues a run and returns its ID for progress andhar_inspectSummarize HTTP requests, status codes, timings and response sizes in a browser network export. The report excludes cookies and authorization headers. HAR JSON up to 32 MiB. URLs are reported without query strings; the source may still contain credentials. No recorded request is rcertificateDecode an X.509 certificate to inspect issuer, subject, validity dates, key algorithm and SHA-256 fingerprint. PEM or DER certificate only, up to 1 MiB. This does not establish chain trust, revocation status or ownership. Private keys are not needed. Queues a run and returns its hex_viewInspect a bounded byte range as hexadecimal and printable ASCII. Choose an offset to examine a file signature or binary structure. Up to 32 MiB input and 64 KiB per displayed range. This is a byte viewer, not code execution or a full binary editor. Queues a run and returns its IDcompressMake images and PDFs fit a size limit. Bounded processing; 32 MiB combined input per operation, at most 50 files. Queues a run and returns its ID for progress and output tracking. Uses cloud allowance.https://sizemyfile.com/mcpSizeMyFile is not in Muse's Connectors list yet, and it needs you to sign in with your SizeMyFile account. Apps that need this kind of sign-in usually cannot be used from a Muse chat yet. Follow it here and we will tell you when it reaches Muse.
At the last check (November 21, 2026, 19:00 UTC) the endpoint was working, answering in 391ms. Over the last 7 days it answered 100% of health checks. It is checked every 15 minutes.
It was screened on November 21, 2026 with the result "screened, no issues found". Screening checks the domain against threat feeds and reads the tools for hidden instructions and requests for passwords or card numbers. It cannot see the server's code, so grant only the access you need.
It is freemium. Check https://sizemyfile.com/agents for current pricing.
It exposes 40 tools, including import_file, list_artifacts, search_tools, prepare_upload. For example, you could ask Muse: "Compress this PDF so it is under 5 MB and give me the download link."
Paste this on your site or README. It always shows the latest check.
<a href="https://musedirectory.ai/connector/sizemyfile"><img src="https://musedirectory.ai/badge/sizemyfile.svg" alt="SizeMyFile on musedirectory.ai" width="236" height="40"></a>