Replace the host bind-mount + containerd exec approach with a per-bot
in-container gRPC server (ContainerService, port 9090). All file I/O,
exec, and MCP stdio sessions now go through gRPC instead of running
shell commands or reading host-mounted directories.
Architecture changes:
- cmd/mcp: rewritten as a gRPC server (ContainerService) with full
file and exec API (ReadFile, WriteFile, ListDir, ReadRaw, WriteRaw,
Exec, Stat, Mkdir, Rename, DeleteFile)
- internal/mcp/mcpcontainer: protobuf definitions and generated stubs
- internal/mcp/mcpclient: gRPC client wrapper with connection pool
(Pool) and Provider interface for dependency injection
- mcp.Manager: add per-bot IP cache, gRPC connection pool, and
SetContainerIP/MCPClient methods; remove DataDir/Exec helpers
- containerd.Service: remove ExecTask/ExecTaskStreaming; network setup
now returns NetworkResult{IP} for pool routing
- internal/fs/service.go: deleted (replaced by mcpclient)
- handlers/fs.go: deleted; MCP stdio session logic moved to mcp_stdio.go
- container provider Executor: all tools (read/write/list/edit/exec)
now call gRPC client instead of running shell via exec
- storefs, containerfs, media, skills, memory: all I/O ported to
mcpclient.Provider
Database:
- migration 0022: drop host_path column from containers table
One-time data migration:
- migrateBindMountData: on first Start() after upgrade, copies old
bind-mount data into the container via gRPC, then renames src dir
to prevent re-migration; runs in background goroutine
Bug fixes:
- mcp_stdio: callRaw now returns full JSON-RPC envelope
{"jsonrpc","id","result"|"error"} matching protocol spec;
explicit "initialize" call now advances session init state to
prevent duplicate handshake on next non-initialize call
- mcpclient Pool: properly evict stale gRPC connection after snapshot
replace (container process recreated); use SetContainerIP instead
of direct map write so IP changes always evict pool entry
- migrateBindMountData: walkErr on directories now counted as failure
so partially-walked trees don't get incorrectly marked as migrated
- cmd/mcp/Dockerfile: removed dead file (docker/Dockerfile.mcp is the
canonical production build)
Tests:
- provider_test.go: restored with bufconn in-process gRPC mock
(fakeContainerService + staticProvider), 14 cases covering all 5
tools plus edge cases
- mcp_session_test.go: new, covers JSON-RPC envelope, init state
machine, pending cleanup on cancel/close, readLoop cancel
- storefs/service_test.go: restored (pure function roundtrip tests)
Split long AI responses into multiple platform messages during streaming
instead of truncating them. The manager counts accumulated delta runes
and opens a new stream when approaching the platform's TextChunkLimit.
Uses a soft/hard limit strategy that prefers splitting at sentence ends
or line breaks over cutting mid-sentence.
- Add pushDelta with soft (75%) / hard (100%) limit and natural break
point detection
- Add splitStream, pushFinalAfterSplit, pushFinalWithChunking helpers
- Fix Discord adapter to use RuneCount Message Length
- Add tests for delta splitting, natural breaks, and final handling
* feat(telegram): use sendMessageDraft for streaming in private chats
Use Telegram Bot API 9.3's sendMessageDraft to stream partial messages
with smooth animation in private chats, replacing the sendMessage +
editMessageText approach. Group/channel chats keep the existing
edit-based streaming.
- Add sendTelegramDraft() for the sendMessageDraft API
- Detect private chats via conversation_type metadata in OpenStream
- Use 300ms throttle for drafts (vs 5s for edits)
- Send permanent messages at tool call boundaries and on final event
- Reset buffer atomically in StreamEventFinal to prevent duplicate
messages when multiple final events fire (one per assistant output)
* test(telegram): improve draft mode test assertions
Add sendTextForTest hook for sendTelegramTextReturnMessage to enable
direct assertion of send calls. Clean up residual unused variables
and replace indirect assertions with explicit mock-based verification.
cni.Remove() failure on stale iptables state blocked the retry
cni.Setup(), leaving bot containers without SNAT/MASQUERADE.
- Ignore cni.Remove() error so retry Setup always runs
- Add global MASQUERADE rule in entrypoints as belt-and-suspenders
Closes#161
- Fix DeleteContainer FAILED_PRECONDITION by cleaning up stopped task
entries before container deletion
- Fix CreateSnapshot leaving container in broken state: commit turns
the active snapshot read-only, so the full cycle (stop → commit →
prepare → delete → recreate → start) is now applied consistently
- Use context.WithoutCancel for atomic container replacement sequences
to prevent cancelled HTTP requests from corrupting container state
- Use dctx for DB operations (recordSnapshotVersion/insertEvent) to
avoid orphan snapshots in containerd without matching DB records
- Restart task + network after snapshot replacement, fixing Exec after
CreateVersion where the container had no running task
- Extract replaceContainerSnapshot helper to deduplicate the prepare →
delete → recreate → start pattern across three call sites
- Move snapshot list data fetching into Manager.ListBotSnapshotData to
encapsulate per-container locking; remove exported LockBot method
- Use UnixNano for snapshot names to avoid second-precision collisions
Allow configuring a custom Telegram Bot API base URL (`apiBaseURL`) per
channel, enabling users behind restricted networks to route requests
through a reverse proxy (e.g. Nginx, Cloudflare Workers).
Both API calls and file downloads respect the configured endpoint.
When omitted, the official https://api.telegram.org is used.
Closes#159
Add mcp-build.sh that compiles the MCP binary and packages it as an
OCI image layer on top of the base rootfs, imported directly into
containerd. Air triggers rebuild on code changes, cleaning stale
containers automatically.
Consolidate dev-only files (Dockerfiles, entrypoint, config, build
script) into devenv/ to separate dev tooling from production artifacts.
Skip image pull when already imported to speed up dev startup.
* feat(models): add per-model probe testing and auto-detect in UI
Move health probes from provider level to model level for precise
testing with real model_id and client_type. Provider test is now a
simple reachability check.
Backend:
- Add POST /models/:id/test endpoint that probes the model's provider
using its actual model_id and client_type
- Add model healthcheck checker for bot health checks (chat/memory/embedding)
- Simplify provider test to reachability-only
Frontend:
- Auto-probe models on mount with status indicator (green/yellow/red dot + latency)
- Auto-probe provider reachability on load and on provider switch
- Fix missing faBolt icon registration
- Manual re-probe via refresh button
Closes#117
* fix(models): increase probe timeout to 15s for slow providers
Some providers (e.g. DashScope) exceed the 5s probe timeout, causing
false-negative "context deadline exceeded" errors. Increase per-probe
timeout to 15s and healthcheck overall timeout to 30s.
* fix(sdk): regenerate exports after merge conflict
Resolve duplicate SDK exports introduced by merge conflict resolution so the web build can compile again while preserving new model probe endpoints.
* feat: add email service with multi-adapter support
Implement a full-stack email service with global provider management,
per-bot bindings with granular read/write permissions, outbox audit
storage, and MCP tool integration for direct mailbox access.
Backend:
- Email providers: CRUD with dynamic config schema (generic SMTP/IMAP, Mailgun)
- Generic adapter: go-mail (SMTP) + go-imap/v2 (IMAP IDLE real-time push via
UnilateralDataHandler + UID-based tracking + periodic check fallback)
- Mailgun adapter: mailgun-go/v5 with dual inbound mode (webhook + poll)
- Bot email bindings: per-bot provider binding with independent r/w permissions
- Outbox: outbound email audit log with status tracking
- Trigger: inbound emails push notification to bot_inbox (from/subject only,
LLM reads full content on demand via MCP tools)
- MailboxReader interface: on-demand IMAP queries for listing/reading emails
- MCP tools: email_accounts, email_send, email_list (paginated mailbox),
email_read (by UID) — all with multi-binding and provider_id selection
- Webhook: /email/mailgun/webhook/:config_id (JWT-skipped, signature-verified)
- DB migration: 0019_add_email (email_providers, bot_email_bindings, email_outbox)
Frontend:
- Email Providers page: /email-providers with MasterDetailSidebarLayout
- Dynamic config form rendered from ordered provider meta schema with i18n keys
- Bot detail: Email tab with bindings management + outbox audit table
- Sidebar navigation entry
- Full i18n support (en + zh)
- Auto-generated SDK from Swagger
Closes#17
* feat(email): trigger bot conversation immediately on inbound email
Instead of only storing an inbox item and waiting for the next chat,
the email trigger now proactively invokes the conversation resolver
so the bot processes new emails right away — aligned with the
schedule/heartbeat trigger pattern.
* fix: lint
---------
Co-authored-by: Acbox <acbox0328@gmail.com>
* feat(search): add Sogou search provider
* fix(search): use new endpoint and API version for sogou
* feat(search): add Serper, SearXNG, Jina, Exa, Bocha, DuckDuckGo search providers
Add six new search provider integrations:
- Serper: Google search via Serper API
- SearXNG: Self-hosted meta search engine
- Jina: Jina AI search API
- Exa: Exa neural search API
- Bocha: Bocha AI web search
- DuckDuckGo: DuckDuckGo HTML search (no API key required)
Each provider includes backend implementation, config schema,
i18n entries, and Vue settings component.
* feat(search): add Yandex search provider
Add Yandex search provider with XML response parsing and
configurable search type (RU/TR/COM).
---------
Co-authored-by: Menci <mencici@msn.com>
Containerd does not auto-prepend "docker.io/" to short Docker Hub names
like "memohai/mcp:latest", causing it to treat "memohai" as a registry
host and fail with EOF. Add NormalizeImageRef() to ensure all image
references are fully qualified before being passed to containerd.
The previous snapshotParentFromLayers manually decompressed layer
blobs to compute diffIDs, which could diverge from the IDs that
containerd's Unpack uses (e.g. gzip vs zstd). Use image.RootFS()
instead — the canonical source containerd itself relies on.
* feat(devenv): add containerized development environment
Replace local-process dev workflow with a fully containerized stack
using docker compose. This enables consistent development across
machines without requiring local Go/Node toolchains or containerd.
- Add Dockerfile.server.dev with containerd + CNI networking support
- Add Dockerfile.web.dev for frontend dev server
- Add server-dev-entrypoint.sh for containerd lifecycle management
- Expand devenv/docker-compose.yml with server, agent, web, migrate
and deps services with proper health checks and dependency ordering
- Update app.dev.toml to use container service names instead of localhost
- Refactor mise.toml dev tasks to drive docker compose workflow
- Support agent_gateway.server_addr in config package for inter-container
communication
* feat(devenv): add hot-reload and registry mirror support
- Add air for Go server hot-reload in dev containers
- Fix agent_gateway host in dev config (0.0.0.0 -> agent)
- Add configurable registry mirror for China mainland users
- Unify MCP image refs via MCPConfig.ImageRef()
* feat(scripts): add China mainland mirror option to install script
Prompt users to opt-in to memoh.cn mirror during installation,
which applies docker-compose.cn.yml overlay and sets registry
in config.toml for MCP image pulls.
BBQ