Word AI
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PYTHONPATH | No | Set the Python path for development | |
| WORD_AI_ENGINE | No | Control offline engine selection: auto, dotnet, python | auto |
| WORD_AI_DOTNET_DLL | No | Override the .NET backend DLL path | |
| WORD_AI_DOTNET_EXE | No | Override the .NET backend executable path | |
| WORD_AI_DOTNET_RID | No | Override the RID detection for native backend loading | |
| WORD_AI_DOTNET_NATIVE_DIR | No | Override the native directory for backend loading | |
| WORD_AI_NPM_USE_SOURCE_BOOTSTRAP | No | Use legacy Python venv bootstrap for npm path (set to 1 to enable) |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {} |
| resources | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| docx_health_checkA | Read-only. Conservative stability report: duplicate tags, complex content controls, fields, tracked changes, comments, table previews and recommended safety policy. |
| docx_inspectC | Read-only. Inspect counts, anchors and optional visible text. |
| docx_package_manifestC | Read-only. DOCX ZIP manifest with part sizes and optional SHA-256 hashes. |
| docx_list_partsA | Read-only. List all package parts with sizes and hashes. |
| docx_part_hashesA | Read-only. Compact DOCX part-hash map for before/after invariance checks. |
| docx_structural_fingerprintA | Read-only. Strong structure fingerprint: package parts, content-control hashes, paragraph hashes, table hashes, outline and health report. |
| docx_mapB | Read-only. Codex-friendly map: health, headings, anchors, paragraph chunks, table summaries and hashes. |
| docx_get_outlineB | Read-only. Heading outline with section paragraph ranges for large-document chunking. |
| docx_outlineA | Read-only alias for docx_get_outline; kept for Codex prompt ergonomics. |
| docx_list_headingsB | Read-only. Heading-only view with previews and hashes. |
| docx_read_heading_sectionC | Read-only. Read one heading section by heading anchor_id or exact heading text. |
| docx_list_anchorsB | Read-only. List stable editing anchors: content controls, headings, bookmarks and paraId paragraphs. |
| docx_read_anchorB | Read-only. Read text behind an anchor_id from docx_list_anchors. |
| docx_search_textB | Read-only. Search visible paragraph text; returns paraId, paragraph index, scope and hashes. |
| docx_list_paragraphsC | Read-only. Paragraph inventory with style, paraId, heading level, scope, preview, hash and complexity. |
| docx_read_paragraphC | Read-only. Read one paragraph by paragraph_index or paraId and return hash/complexity. |
| docx_list_content_controlsB | Read-only. List content controls with tag, alias, id, preview, text hash and complexity. |
| docx_read_content_controlA | Read-only. Read content-control text by w:tag. Preferred primitive for safe edits. |
| docx_extract_plain_textA | Read-only. Extract visible body text for coarse review; never use to regenerate DOCX. |
| docx_export_plain_textC | Write-sidecar only. Export visible body text to .txt; does not modify DOCX. |
| docx_list_tablesB | Read-only. List table dimensions, scope tags, complexity and preview rows. |
| docx_read_tableB | Read-only. Read a table matrix by table_index and optional content-control scope_tag. |
| docx_read_table_cellC | Read-only. Read a single table cell and return hash for write preconditions. |
| docx_export_table_csvB | Write-sidecar only. Export table to CSV; does not modify DOCX. Uses default sidecar naming when out_path is omitted. |
| docx_table_to_csvC | Write-sidecar only. Export table to CSV using default naming when out_path is omitted. |
| docx_list_stylesA | Read-only. List Word styles from styles.xml for style-aware generation. |
| docx_list_numberingB | Read-only. Inspect numbering.xml numId/abstractNum mappings. |
| docx_list_sectionsB | Read-only. Inspect section properties, margins, page size and header/footer refs. |
| docx_list_fieldsB | Read-only. List field instructions such as TOC, REF, PAGEREF and SEQ. |
| docx_list_imagesC | Read-only. List drawing/pict objects with relationship IDs and alt metadata. |
| docx_list_hyperlinksA | Read-only. List hyperlinks with relationship targets, preview and hashes. |
| docx_list_commentsC | Read-only. Extract comments.xml previews, author/date/id and hashes. |
| docx_extract_commentsC | Read-only. Extract full comment metadata and text. |
| docx_list_bookmarksC | Read-only. List bookmarks and paths for cross-reference preservation. |
| docx_list_headers_footersB | Read-only. Inspect header/footer parts and visible text without modifying them. |
| docx_list_notesC | Read-only. Inspect footnotes/endnotes when present. |
| docx_list_revisionsA | Read-only. List tracked-change nodes with previews and hashes. |
| docx_assess_patchsetB | Read-only. Resolve PatchSet targets and report risks, touched objects and precondition gaps. Defaults to the .NET Open XML backend when available. |
| docx_plan_patchsetD | Read-only. Alias-grade PatchSet planning assessment. Defaults to the .NET Open XML backend when available. |
| docx_preflight_patchsetB | Write-temp. Dry-run PatchSet to a temporary copy, validate, and remove output unless keep_output=true. Defaults to the .NET Open XML backend when available. |
| docx_dry_run_patchsetC | Write-temp. Same as preflight; useful for explicit Codex dry-run stage. Defaults to the .NET Open XML backend when available. |
| docx_write_indexC | Write-sidecar only. Write .wordai index JSON; does not modify DOCX. |
| docx_backupB | Write-sidecar only. Create timestamped backup; does not modify original DOCX. |
| docx_restore_backupC | Write. Restore a backup to target_path. |
| docx_rollbackD | Write. Restore from backup and optionally back up replaced file first. |
| docx_apply_patchsetC | Write. Apply constrained PatchSet to a new DOCX plus audit JSON; validation gates final commit. Defaults to the .NET Open XML backend when available. |
| docx_validateB | Read-only. Validate structural invariants. Supply touched_* for intentional edits. Defaults to the .NET Open XML backend when available. |
| docx_compare_structureC | Read-only. Same validation report phrased as structural comparison. Defaults to the .NET Open XML backend when available. |
| docx_text_diffB | Read-only. Unified visible-text diff for human review after validation. |
| word_session_listB | Read-only. List active Office.js taskpane sessions registered by an open Word document. |
| word_session_snapshotA | Read-only. Return the latest content-control snapshot for an open Word session. If session_id is omitted, uses the most recently active session. |
| word_session_refreshB | Read-only. Ask the Office.js taskpane to refresh its open-document content-control snapshot. |
| word_session_read_content_controlB | Read-only. Ask the Office.js taskpane to read content-control text from the currently open Word document by tag. |
| word_session_preview_patchsetB | Read-only. Ask the Office.js taskpane to validate and preview a PatchSet against the currently open Word document without modifying it. |
| word_session_apply_patchsetC | Write. Apply a supported content-control PatchSet to the currently open Word document through Office.js, with hash preconditions, audit, and rollback PatchSet. |
| word_session_wrap_selectionC | Write. Ask the Office.js taskpane to wrap the current Word selection in a content control with a stable tag/title. |
| word_session_rollbackC | Write. Roll back a previous word_session_apply_patchset command by applying its generated rollback PatchSet to the open Word document. |
| word_session_command_statusA | Read-only. Return the current status/result/error for a queued Office.js session command. |
| officecli_view_htmlC | Optional OfficeCLI auxiliary read. Render a DOCX to a bounded HTML snapshot for visual evidence. Does not modify the DOCX. |
| officecli_view_screenshotC | Optional OfficeCLI auxiliary sidecar export. Render a DOCX screenshot/PNG to output_path. Does not modify the DOCX. |
| officecli_view_issuesC | Optional OfficeCLI auxiliary read. Run view issues with JSON output for formatting/content/structure evidence. |
| officecli_queryB | Optional OfficeCLI auxiliary read. Run CSS-like query with --json. Use only for inspection, never mutation. |
| officecli_validateC | Optional OfficeCLI auxiliary read. Validate the DOCX with OfficeCLI --json as extra evidence after Word AI validation. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/flyfish-dev/word-ai'
If you have feedback or need assistance with the MCP directory API, please join our Discord server