so the v6 hook stops re-creating its store every render on a new chat
(which wiped the optimistic user message + streamed deltas, so nothing
showed until the turn finished). Also send X-Accel-Buffering:no + flushHeaders.
- context: client sends the currently-open page {id,title}; the system prompt
tells the agent which page 'this page' refers to (it reads it via its
CASL-scoped getPage tool; id is prompt-context only, no server-side fetch).
- embeddings: make page_embeddings.embedding dimension-agnostic (drop the
HNSW index + ALTER to vector), remove the hard 1536 guard, filter search by
model_dimensions — so 3072-dim (and any) models index instead of being
skipped. Seq-scan <=> search (wiki scale); existing pages reindex on next edit.
- openai provider: use .chat() (Chat Completions) instead of the default callable
(Responses API), which gateways reject on multi-turn -> 400.
- updateAiProviderSettings: assemble settings.ai.provider via jsonb_build_object
with ::text-cast bound params + jsonb_typeof self-heal (postgres.js was
double-encoding it into an array; the ::text cast avoids 'could not determine
data type of parameter').
- chat agent: drop the hard maxOutputTokens cap (truncated complex tool calls);
keep a tiny cap only on the test-connection ping.
- testConnection + chat stream: surface the real provider error (statusCode+message)
to logs and the UI instead of generic masks; never log the API key.
- chat UI: typing indicator, incremental streaming render, tool 'running' status, Stop.
Also bundled (prior uncommitted ai-chat work):
- history 'AI agent' provenance badge; vector RAG (pgvector image + page_embeddings
+ AI_QUEUE indexer + space-scoped semanticSearch); external MCP servers backend
(@ai-sdk/mcp client, SSRF IP-pinning, encrypted headers, admin CRUD/Test);
yjs duplicate-instance fix via pnpm patch (single CJS instance server-side).
- Add reversible write tools to the per-user agent toolset (page create/update/
move/soft-delete; comment reply + resolve), exposed under the user's JWT and
enforced by Docmost CASL; no permanent/force delete (D3).
- Non-spoofable agent provenance: sign actor/aiChatId into the access and collab
tokens (TokenService), propagate via jwt.strategy onto the request, and set
pages.last_updated_source/last_updated_ai_chat_id on REST create/update/move and
comments.created_source/resolved_source/ai_chat_id.
- packages/mcp: add an optional getCollabToken provider (content-edit provenance)
and guard against empty tokens; service-account /mcp path unchanged.
Frontend:
- Admin 'AI / Models' settings section: provider/model/embedding/base URL, a
write-only API key field, system prompt, and Test connection.
- AI chat panel (useChat + DefaultChatTransport): conversation list, streamed
messages, tool-call action log and page citations; header entry point gated on
settings.ai.chat.
Compile-verified (server nest build + client tsc/vite); not yet live-tested.
Known gaps: history 'AI agent' badge (C3), vector RAG (D), external MCP (E);
chat tool-card citation links pending a fix.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
WIP checkpoint of the gitmost AI-chat backend (plan stages A + B1 + B3a).
The agent acts under the requesting user's JWT (Docmost CASL enforces page
access); the external service-account /mcp endpoint is untouched.
LLM provider config (A2-A4):
- integrations/crypto: AES-256-GCM SecretBoxService (key derived from APP_SECRET,
per-record salt/iv; clear error on rotation instead of crashing).
- ai_provider_credentials table/repo/types: encrypted API key stored outside
workspace settings/baseFields, write-only (never returned by any endpoint).
- integrations/ai: per-workspace AI SDK v6 provider driver (openai/gemini/ollama),
admin-gated GET(masked)/PATCH(write-only key)/Test endpoints; settings.ai.provider
holds non-secret config incl. systemPrompt. Removed unused AI_* env getters (DB is
the single source of truth).
Chat module (A1, A5-A8):
- ai_chats/ai_chat_messages repos (workspace-scoped, soft-delete, tsv never selected).
- core/ai-chat: CRUD + POST /ai-chat/stream (Fastify hijack + AI SDK v6
pipeUIMessageStreamToResponse, abort on disconnect, persist user/assistant msgs).
- Agent loop: streamText + stepCountIs(8); read tools searchPages/getPage via a
per-request DocmostClient over loopback REST under the user's minted access token.
- Gate settings.ai.chat (+ 503 when provider unconfigured); buildSystemPrompt with a
non-removable safety/anti-prompt-injection framework. Per-user rate limit.
Per-user auth (B1):
- @docmost/mcp DocmostClient gains an additive getToken variant (carry a user JWT,
re-fetch on 401) and exports DocmostClient; the email/password service-account path
(external /mcp, stdio) is unchanged.
Agent-edit provenance backbone (B3a):
- Migration: pages/page_history (last_updated_source, last_updated_ai_chat_id) and
comments (created_source, ai_chat_id, resolved_source).
- Signed actor/aiChatId claim in the collab token; onAuthenticate propagates it,
onStoreDocument writes it with a sticky agent marker, saveHistory copies it.
Migrations auto-run on boot (additive). Write tools, frontend, RAG and external MCP
servers are not in this checkpoint.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the removed enterprise EE MCP (private apps/server/src/ee submodule,
license-gated /mcp route) with our docmost-mcp, vendored as an isolated ESM
workspace package and served by the server over HTTP — no enterprise license.
Backend:
- Add packages/mcp (@docmost/mcp): vendored docmost-mcp refactored into a
side-effect-free createDocmostMcpServer() factory (38 tools preserved),
stdio entry kept in stdio.ts, Streamable-HTTP session manager in http.ts.
- Add apps/server McpModule: @Post/@Get/@Delete('mcp') (served at /mcp via the
existing global-prefix exclude), @SkipTransform + reply.hijack to bridge raw
Fastify req/res into the SDK transport. The module dynamically imports the
ESM-only package from CommonJS via a Function-indirected import resolved with
require.resolve + file:// URL. Gated by the workspace ai.mcp toggle, a
service-account (MCP_DOCMOST_EMAIL/PASSWORD/API_URL) and optional MCP_TOKEN;
per-session idle eviction (MCP_SESSION_IDLE_MS).
- Drop the enterprise license check on mcpEnabled in workspace.service.
- Dockerfile: copy packages/mcp into the production image.
- .env.example: document MCP_DOCMOST_*, MCP_TOKEN, MCP_SESSION_IDLE_MS.
Frontend:
- Recreate the community "AI & MCP" workspace-settings panel (mcp-settings.tsx):
admin-only toggle on settings.ai.mcp with optimistic update, copyable
${APP_URL}/mcp URL; wired into workspace-settings page. Reuses existing i18n.
Fixes:
- Pin packages/mcp tiptap deps to 3.20.4 (matching the client) and inline
getStyleProperty, preventing a duplicate @tiptap/core@3.26.1 from leaking into
the client editor via pnpm shamefully-hoist (was breaking apps/client tsc).
Add comment resolve/re-open as a community feature, written from scratch on top
of the infrastructure already present in the community codebase: the
resolved_at/resolved_by_id columns, the COMMENT_RESOLVED notification job, the
resolveCommentMark collaboration handler, the commentResolved websocket event,
the comment service/types and the Open/Resolved tabs. No Enterprise-Edition code
is reused and there is no EE feature gating — resolving is available to anyone
who can comment.
Backend:
- add POST /comments/resolve (ResolveCommentDto) guarded by validateCanComment;
reject resolving replies
- add CommentService.resolveComment: set/clear resolvedAt/resolvedById, sync the
inline comment mark via collaboration handleYjsEvent, queue
COMMENT_RESOLVED_NOTIFICATION (only when another user resolves), emit the
commentResolved websocket event and write a resolve/reopen audit log
Frontend:
- add useResolveCommentMutation with optimistic update + rollback
- add ResolveComment toggle button
- wire the resolve button and menu item into comment-list-item / comment-menu,
gated on canComment for parent comments
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Port the compact page-tree indentation tweak from the docmost-app
WebKit wrapper (previously applied via injected CSS) into the source.
- doc-tree.tsx: change the default `indentPerLevel` prop from 16 to 8,
giving an 8px step per nesting level for a more compact page tree.
No callers override the value, so the single default change applies
everywhere DocTree is rendered. EE-hiding injections from docmost-app
were skipped: those surfaces are already removed in this community fork.
Strip the proprietary client EE so the fork ships a clean community/AGPL
edition, mirroring Forkmost. Delete apps/client/src/ee (201 files) and
packages/ee, and patch every consumer that imported from @/ee/*.
- gate-out EE features (useHasFeature -> false): API keys, SSO, MFA, SCIM,
audit logs, AI / AI-chat, templates, page permissions, page verification,
comment resolution, trash retention, viewer comments
- drop cloud/billing/trial/entitlement/posthog flows; sign-in is now
email+password only (no SSO/LDAP/cloud)
- remove EE routes from App.tsx and EE entries from sidebars/settings nav
- restore the community page-share button (ShareModal) that the EE
PageShareModal used to provide
- remove the dead "Attachments" search filter, dead MFA navigation and
orphaned route constants
Client type-checks clean; full `pnpm build` is green for all three projects.
Remove the private apps/server/src/ee git submodule (github.com/docmost/ee)
and the now-empty .gitmodules so that `git clone --recurse-submodules` and CI
checkout no longer fail with 404. The server loads EE only via guarded runtime
require(), so the build succeeds without it (community edition).
Rewrite .github/workflows/release.yml for the fork:
- drop the GitHub App token step and `submodules: recursive` checkout
- publish to GHCR (ghcr.io/vvzvlad/gitmost) via the built-in GITHUB_TOKEN
instead of Docker Hub (docmost/docmost) — no extra secrets required
- add `packages: write` permission and an IMAGE env var
- log in as github.repository_owner; rename release tarballs to gitmost-*
Repoint the Dockerfile image source label to the fork.
* feat(editor): add alt text support for images
* feat: extend alt text support to videos and diagrams
---------
Co-authored-by: Philipinho <16838612+Philipinho@users.noreply.github.com>
* Better trash
I recently lost a bunch of time editing and searching for pages that were actually in the Trash. Docmost intentionally tries to not link to Trashed pages, but the url of that Trashed page and any inbound links still work. This makes it clearer when a page you are interacting with is in the Trash.
- /trash
- Refactored banner into `trash-banner.tsx`
- Refactored "Restore" modal into `use-restore-page-modal.tsx`
- Page (when isDeleted)
- Add: `trash-banner.tsx`
- Add breadcrumbs: `Parent / Child / Page (Deleted)`
- Change: Deleted Pages are read-only
- Replace "Move to Trash" with "Restore" in page menu (invokes `use-restore-page-modal`)
I tried very hard to keep this simple and re-use existing translation strings wherever possible.
* cleanup
---------
Co-authored-by: Philipinho <16838612+Philipinho@users.noreply.github.com>
The header edit/read toggle now controls only the current session's mode
without saving it as the user's preference. The saved preference (set in
profile settings) is applied once on initial load and sticks across page
navigations within the session, so navigating to a new page no longer
resets the mode mid-session.
Fixes#1693
* fix(editor): hide transclusion borders and reset spacing in read-only mode
* feat(share): add full width toggle for shared pages
* feat(share): support resizing sidebar on shared pages
* fix: auto redirect if there is only one SSO provider.
- fix tighten sso redirect
- fix share tree margin
* sync
* package overrides
* feat(tree): replace react-arborist with custom tree implementation
* feat(tree): keyboard arrow navigation between rows
* feat(emoji-picker): focus search input on open
* refactor(emoji): switch to @slidoapp/emoji-mart fork for accessibility
* feat(tree): Home/End and typeahead keyboard navigation
* feat(tree): roving tabindex and * to expand sibling subtrees
* feat(tree): Space activation and ARIA refinements
* fix(tree): move treeitem role to focusable row + aria-current
* feat(editor): show emoji name in suggestion list
Replace the fixed-column emoji grid with a vertical list that displays
each emoji alongside its :shortcode: name. This makes the picker more
discoverable—users can see and learn shortcodes without prior knowledge.
Changes:
- EmojiList: switch from SimpleGrid/ActionIcon to UnstyledButton list
rows showing emoji glyph + monospace 🆔 label
- Navigation simplified to ArrowUp/ArrowDown (list has no columns)
- Results capped at 8 items for a focused, scannable dropdown
- CSS module: rename menuBtn -> menuItem, tighten padding
* feat(editor): replace SearchIndex with name/id includes search
Port the exact search algorithm from the original extension:
- Build a flat index from @emoji-mart/data: { id, name (lowercase), native }
- Filter with name.includes(q) || id.includes(q) — predictable, no
keyword indirection
- Results capped at 5 (same as extension)
- Frequently-used emojis (sorted by usage) shown when query is empty
- Remove emoji-mart init() / SearchIndex / getEmojiDataFromNative
dependencies; index is built lazily and cached in memory
- Remove unused GRID_COLUMNS constant
* feat(editor): emoji picker with browse and search modes
When the query is empty the picker shows a category bar with 8 tabs
(people, nature, food…) and a scrollable emoji grid. Typing after ':'
switches to a compact list that shows the glyph and :shortcode: side by
side, making it easy to discover emoji names while you type.
- Category data is loaded lazily from @emoji-mart/data and cached, so
opening the picker more than once has no overhead
- Grid keyboard nav: arrow keys move by cell/row, Enter picks
- List keyboard nav: up/down through results, Enter picks
- Mouse hover syncs the keyboard selection index in both modes
- incrementEmojiUsage tracks picks so frequently used ones bubble up
in future sessions
* fix(editor): polish emoji picker copy and loading
* feat: add emoji to slash command
* Add keyboard support to emoji group navigation
---------
Co-authored-by: Philipinho <16838612+Philipinho@users.noreply.github.com>
