Resolve intent · plan fit · stay user-controlled

Let AI start from the user's file job instead of guessing a FileNest tool name.

FileNest now separates natural-language routing → fit planning → action. Intent Router v1.0 deterministically maps one task sentence to a published tool, workflow or no-match without calling an external model or storing the sentence; fixed answer clusters can then use AI Task Planner v1.2 for Good Fit, Not Fit, evidence and the next step. Real file selection, preview, processing and download remain with the user.

Start an actual file task Open intent router contract AI Task Planner WebMCP Web task contract Human task router

Arrived from an AI assistant? Start with the file job

You do not need to install MCP to use FileNest.

Choose an existing human-operated tool below. You select files, confirm settings, review the output and decide when to download. The read-only remote MCP never processes your files inside a chat.

Online Batch File Renamer

Preview and create renamed copies in bulk without overwriting the source files.

Basic guest access available; check the live tool's limits.

Open the actual tool

Batch Image Compression & Size Optimization

Convert, resize, compress or enhance JPG, PNG and WebP images in the browser.

Sign-in required; the live tool controls quotas and limits.

Open the actual tool

Merge, Split & Create PDFs in Batches

Merge PDFs, split/extract pages, or create PDFs from JPG/PNG images in the browser.

Sign-in required; the live tool controls quotas and limits.

Open the actual tool

Batch CSV, TSV & JSON Conversion

Parse and convert CSV, TSV and JSON locally in the browser.

Sign-in required; the live tool controls quotas and limits.

Open the actual tool

Browse all published tools → · Check current Free, eligible 24H Pass and Pro terms →

WEB TASK CONTRACT 1.0

Declare what can be done and what must never happen silently.

Contract last updated 2026-08-30. Current live tool behavior always outranks stale cached descriptions.

File selection stays with the user

AI should not bypass the browser's local file-selection permission.

Routing and planning are not execution consent

Neither routing nor planning reads files or runs a task; subsequent action still requires a separate call and live-page confirmation.

No destructive automation

No automatic deletion or source-file overwrite, and routing, planning, clicks or prefill are never treated as task success.

WEBMCP SITE TOOLS 1.6

From one sentence to a real route, then to separately confirmed action.

WebMCP capability manifest last updated 2026-08-31. WebMCP is currently an experimental web standard; FileNest describes 10 bounded site tools through the browser's document.modelContext API, with both natural-language routing and the AI Task Planner remaining read-only decision layers.

list_filenest_capabilities

Read-only listing of 13 current task entry points, purposes, processing boundaries and access requirements without reading files.

resolve_filenest_intent

Takes a task description of up to 500 characters and uses FileNest's deterministic rules to return a published task, workflow or no-match. It calls no external or generative model, stores no task sentence, accesses no files and does not navigate; match strength is not a statistical probability.

plan_filenest_task

Read-only fit planning over fixed FileNest answer clusters. Returns Good Fit, Not Fit, first-party evidence, a recommended real task and next step. Supports no-match and does not navigate, prefill, access files or execute processing.

open_filenest_task

Navigates only from a fixed task-ID allowlist to real same-origin FileNest pages. It accepts no arbitrary URL and performs no file processing.

open_filenest_workflow

Navigates only from the current published workflow-ID allowlist to the matching workflow anchor. It does not read or transfer files, silently carry one step's output into the next tool, or auto-advance progress.

prepare_batch_rename

Stores bounded core rules such as prefix, suffix, find/replace, extension preservation and numbering in same-origin sessionStorage, then opens the free local renamer. It selects, reads and renames no files, excludes regex and CSV mapping, and expires within one hour.

prepare_data_conversion

Stores only the target output format (JSON / CSV / TSV) and JSON pretty-print setting in same-origin sessionStorage, then opens the local data converter. It does not select, read, parse or convert files, accepts no arbitrary delimiter or remote URL, and expires within one hour.

prepare_pdf_workflow

Stores only published PDF merge, split/extract or images-to-PDF mode plus the relevant page range, per-page split or A4 / Letter / match-image setting in same-origin sessionStorage. It selects no files, reads no PDFs, never processes or downloads, expires within one hour, and does not add encrypted-PDF, PDF OCR or PDF-to-Office support.

prepare_web_image_delivery

Stores bounded format, quality, scale, crop, light-adjustment and naming settings in same-origin sessionStorage, then opens Web Image Delivery. It selects no files, never auto-runs, and expires within one hour.

get_filenest_safety_boundary

Read-only return of file-selection, processing-confirmation, final-download and non-destructive-action boundaries.

Intent → plan → action, never one sentence = automatic execution

resolve_filenest_intent only answers where a job should go and returns no-match for unsupported, conflicting or weak signals. Fixed answer clusters can then call the AI Task Planner for fit and evidence. Only a later separate navigation or Safe AI Prefill action changes page state.

Machine-readable Intent Router contract →
Machine-readable AI Task Planner contract →

WebMCP is not a remote MCP server

These tools exist only in the current webpage context of a WebMCP-aware browser. Natural-language routing and the Task Planner are read-only decision layers; Safe AI Prefill stores only bounded settings or rename rules, stores no file content and never starts processing, renaming, data conversion or PDF work automatically. Browser WebMCP does not expose a remote execution API. FileNest also offers a separate read-only remote MCP task-discovery endpoint, described below; it cannot read or convert user files.

External AI connection · read-only

Connect an MCP-compatible AI client to discover real FileNest tools.

This separate remote MCP endpoint lists FileNest's published file tasks and exact tool URLs without guessing a tool name. It cannot read user files, run conversions or make payments. It is distinct from the in-browser WebMCP tools described above.

Read-only MCP server URL · for compatible client configurationhttps://filenest-pages.pages.dev/mcp

Transport: Streamable HTTP. Tools: list_filenest_tasks and get_filenest_task. Both expose public first-party task descriptions only. This is not a browser landing page; a direct browser GET returns 405. Availability must be verified with a real client handshake and public HTTP checks.

Check fit and exclusions before linking to a real user-operated tool. The 24H Pass applies only to eligible standard processing; live Pricing controls terms. An endpoint deployment does not imply third-party listing or guaranteed AI recommendations.

Read the real-task catalog (JSON) →No MCP client? Describe your file task →Check live plans and entitlements →

Connect from an MCP-compatible AI client (opt-in)

These instructions connect to the already public read-only task finder only. They never let an AI read, upload or modify user files or buy a plan. Verify a real client handshake and the two read-only tools before relying on the connection; no provider listing or recommendation is implied.

Claude

In a Claude account with custom remote connectors, open Customize → Connectors → Add custom connector and enter the full HTTPS endpoint above. Enable the connector in the conversation. Team-managed workspaces may require an owner to add it first.

Claude official connection guide

Cursor

Add this JSON to your personal ~/.cursor/mcp.json or a project .cursor/mcp.json. Then confirm the client really lists list_filenest_tasks and get_filenest_task.

{
  "mcpServers": {
    "filenest": {
      "url": "https://filenest-pages.pages.dev/mcp"
    }
  }
}

Review and add FileNest in Cursor Cursor official MCP install-link guide Cursor official MCP setup

VS Code

If your VS Code version supports remote MCP, place this HTTP entry under servers in .vscode/mcp.json. Review the tool list and trust prompt; connecting alone does not send local files to FileNest.

{
  "servers": {
    "filenest": {
      "type": "http",
      "url": "https://filenest-pages.pages.dev/mcp"
    }
  }
}

VS Code official MCP setup

ChatGPT custom MCP apps require an eligible account or workspace and any required admin/developer permissions. This page does not add FileNest to the public ChatGPT app directory. If a client cannot connect, the real browser-operated file tools above remain available. Official availability requirements

Real web tasks

Every entry point has an explicit success definition and recovery path.

web-batch-file-renamer

Online Batch File Renamer

Open live tool

Preview and create renamed copies in bulk without overwriting the source files.

Inputs

Ordinary files selected from the user's device; multiple files supported

Outputs

ZIP containing renamed copies

Access & processing

Guest basic use allowed

Browser local: Filename rules, ordering, conflict checks, Pro CSV mapping parsing and device-workset checks run in the current browser; once download is confirmed, that preview is frozen as the execution snapshot and the workset is locked until ZIP packaging finishes, so ZIP contents, success accounting and download records use the same snapshot; invalid CSV mappings are not partially applied.

Success means

All proposed names pass conflict and unsafe-name checks and the renamed-copy ZIP is generated successfully.

If it fails

Adjust rules when duplicate or invalid names are detected
Reduce batch size if browser packaging cannot finish

web-sequential-file-renamer

Rename Files in Sequential Order

Open live tool

Preview and create renamed copies in bulk without overwriting the source files.

Inputs

Ordinary files selected from the user's device; multiple files supported

Outputs

ZIP containing renamed copies

Access & processing

Signed-in account required

Success means

All proposed names pass conflict and unsafe-name checks and the renamed-copy ZIP is generated successfully.

If it fails

Adjust rules when duplicate or invalid names are detected
Reduce batch size if browser packaging cannot finish

web-reverse-number-file-renamer

Reverse Number File Renamer

Open live tool

Preview and create renamed copies in bulk without overwriting the source files.

Inputs

Ordinary files selected from the user's device; multiple files supported

Outputs

ZIP containing renamed copies

Access & processing

Signed-in account required

Success means

All proposed names pass conflict and unsafe-name checks and the renamed-copy ZIP is generated successfully.

If it fails

Adjust rules when duplicate or invalid names are detected
Reduce batch size if browser packaging cannot finish

web-custom-order-file-renamer

Custom Order File Renamer

Open live tool

Preview and create renamed copies in bulk without overwriting the source files.

Inputs

Ordinary files selected from the user's device; multiple files supported

Outputs

ZIP containing renamed copies

Access & processing

Signed-in account required

Success means

All proposed names pass conflict and unsafe-name checks and the renamed-copy ZIP is generated successfully.

If it fails

Adjust rules when duplicate or invalid names are detected
Reduce batch size if browser packaging cannot finish

web-image-batch-converter

Image Batch Converter

Open live tool

Convert, resize, compress or enhance JPG, PNG and WebP images in the browser.

Inputs

JPG / JPEG
PNG
WebP

Outputs

JPG, PNG or WebP images
ZIP download for multiple outputs

Access & processing

Signed-in account required

Browser local: These image-processing paths run locally in the browser, including full-resolution detail sharpening and source-container preflight; generated images are downloaded directly by the user.

Success means

Target images are generated and dimensions, byte size and warnings can be checked per image before download.

If it fails

Use JPG, PNG or WebP input
Reduce batch size if the device runs low on memory or processing fails

web-image-resizer

Batch Image Resize, Crop & Zoom

Open live tool

Convert, resize, compress or enhance JPG, PNG and WebP images in the browser.

Inputs

JPG / JPEG
PNG
WebP

Outputs

JPG, PNG or WebP images
ZIP download for multiple outputs

Access & processing

Signed-in account required

Success means

Target images are generated and dimensions, byte size and warnings can be checked per image before download.

If it fails

Use JPG, PNG or WebP input
Reduce batch size if the device runs low on memory or processing fails

web-image-enhancer

Batch Image Clarity & Color Enhancer

Open live tool

Convert, resize, compress or enhance JPG, PNG and WebP images in the browser.

Inputs

JPG / JPEG
PNG
WebP

Outputs

JPG, PNG or WebP images
ZIP download for multiple outputs

Access & processing

Signed-in account required

Success means

Target images are generated and dimensions, byte size and warnings can be checked per image before download.

If it fails

Use JPG, PNG or WebP input
Reduce batch size if the device runs low on memory or processing fails

web-compress-images-online

Batch Image Compression & Size Optimization

Open live tool

Convert, resize, compress or enhance JPG, PNG and WebP images in the browser.

Inputs

JPG / JPEG
PNG
WebP

Outputs

JPG, PNG or WebP images
ZIP download for multiple outputs

Access & processing

Signed-in account required

Success means

Target images are generated and dimensions, byte size and warnings can be checked per image before download.

If it fails

Use JPG, PNG or WebP input
Reduce batch size if the device runs low on memory or processing fails

web-pdf-batch-converter

Merge, Split & Create PDFs in Batches

Open live tool

Merge PDFs, split/extract pages, or create PDFs from JPG/PNG images in the browser.

Inputs

PDF
JPG / JPEG
PNG

Outputs

PDF
ZIP download for multiple outputs

Access & processing

Signed-in account required

Browser local: Supported PDF merge, split, page extraction and JPG/PNG-to-PDF operations, device-workset checks and multi-result safe-packaging decisions run in the browser. Before a confirmed run can start, the visible file order and processing settings must match the synchronous intent state and every input must have complete operation-preview evidence; page and output counts are recomputed from that confirmed preview and frozen into one execution snapshot before device-capacity recheck, synchronous execution lock and quota reservation, and processing, success accounting and result registration use only that snapshot. Before credits are reserved, images-to-PDF checks the actual PNG/JPEG signature and performs a decode probe with the same pdf-lib path used for final embedding; damaged, disguised or incompletely previewed inputs are blocked before quota reservation. If local results are ready but account settlement cannot be confirmed, downloads remain available while new PDF runs and setting mutations are locked; Retry account sync settles only the existing job and does not reprocess files or reserve quota again. Multi-result ZIP creation first freezes the result set and synchronously locks packaging; split and extracted output names derived from source filenames are cleaned for cross-platform invalid characters and Windows reserved names before collision numbering.

Success means

The selected PDF operation completes and downloadable PDF output is generated.

If it fails

Correct invalid page ranges
Repair or unlock damaged/encrypted PDFs elsewhere first

web-csv-json-converter

Batch CSV, TSV & JSON Conversion

Open live tool

Parse and convert CSV, TSV and JSON locally in the browser.

Inputs

CSV
TSV
JSON object or array of objects

Outputs

CSV
TSV
JSON

Access & processing

Signed-in account required

Browser local: CSV/TSV/JSON input normalization, device-workset checks, strict quote-state parsing, preview, output serialization, safe-packaging decisions and downloadable output-name cleanup run in the current browser. Explicitly selected supported files are preserved in selection order instead of being silently deduplicated by filename, size and lastModified metadata; even identical metadata selections remain distinct inputs, while deterministic output numbering prevents overwrites. When conversion is confirmed, that run freezes its file set, target format, JSON pretty-print setting and count metadata as one execution snapshot, then uses a synchronous execution lock through local parsing, quota reservation and result registration; only files that parse successfully inside that snapshot are reserved against quota. If local results are ready but account job settlement cannot be confirmed, the results remain downloadable while new conversions and setting mutations are locked; Retry account sync settles only the existing job and does not reconvert files or reserve quota again. Multi-result ZIP creation first freezes the result set and synchronously locks packaging so repeated triggers or setting changes cannot split the downloaded archive from the visible results. Unclosed quotes, illegal quotes inside unquoted fields and illegal characters after closing quotes are rejected before conversion instead of silently rewriting data; result names derived from source filenames avoid common cross-platform invalid characters and Windows device-reserved names before batch collision numbering.

Success means

Input parses successfully and the target serialized format is generated with an output preview available before download.

If it fails

Fix unclosed CSV/TSV quoted fields
Reshape JSON into an object or array of objects

web-text-to-document-converter

Text to Document Converter

Open live tool

Generate common document formats from user-entered text in the browser with layout preview first.

Inputs

User-entered or pasted text
Optional JPG / PNG / WebP background

Outputs

DOCX
PDF
TXT
Markdown
HTML

Access & processing

Signed-in account required

Browser local: Entered text, preview, built-in templates, custom backgrounds, document generation and downloadable output-name sanitization continue in the current browser; output names avoid common cross-platform invalid characters and Windows device-reserved names.

Success means

After text, layout and background are previewed, the requested document is generated successfully in the browser.

If it fails

Simplify the background or content complexity if generation fails
Recheck layout and filename before download

web-image-to-text-ocr

Image to Text OCR & Document Export

Open live tool

Recognize text from JPG, PNG or WebP images and require human review before export.

Inputs

JPG / JPEG
PNG
WebP

Outputs

Editable recognized text
DOCX
PDF
TXT
Markdown
HTML

Access & processing

Signed-in account required

Browser local: Local Privacy mode uses Tesseract.js in the current browser; selected images are not uploaded for OCR recognition. OCR runtime assets or language data may download on first use.Transient cloud: After explicit user confirmation, High Accuracy mode transiently submits selected JPG, PNG or WebP images to the FileNest OCR Worker, which uses Azure Document Intelligence prebuilt-read for the recognition request. The third-party OCR provider processes that request under its service; FileNest's current implementation does not write OCR source images to R2, KV or the FileNest database.

Success means

Recognition returns editable text; the user must review it before deciding whether to export a document.

If it fails

Use a clearer, straighter image
Switch between Local Privacy and High Accuracy according to the privacy requirement

web-duplicate-file-finder

Content-Level Duplicate File Finder

Open live tool

Compare file fingerprints in the browser to find byte-identical duplicate files.

Inputs

Ordinary files selected from the user's device; multiple files supported

Outputs

Duplicate groups
Optional CSV report

Access & processing

Guest basic use allowed

Browser local: Files are prefiltered by byte size and then compared using chunked SHA-256 fingerprints computed in the browser; an unreadable candidate is isolated without failing the remaining candidates. Duplicate-file CSV reports are also generated locally; if the user stops the scan or candidates are unreadable, both the UI and exported report explicitly mark the result as incomplete and the CSV records scan status, processed candidates, total candidates and unreadable count.

Success means

Files complete byte-size prefiltering and SHA-256 fingerprint comparison and byte-identical groups are shown.

If it fails

Reduce batch size if a large comparison stresses the browser
The user must make any final deletion decision on their own device

Not a public remote file API

Browser agents can resolve intent, plan and collaborate more reliably while file permissions and final decisions stay in the page.

FileNest currently prioritizes reliable AI discovery, task routing, fit decisions, planning and bounded on-page actions, then uses real successful-task and paid-demand signals to decide whether deeper automation is justified. Remote file processing is not silently enabled by the Intent Router, WebMCP, the Task Planner or Safe AI Prefill.

Verify the boundaryTrust Center & data handling →Tests & first-party evidence →Supported formats & limits →AI intent router contract →AI task planner contract →WebMCP capability manifest →Machine-readable task contract →