a5414d83bd
Normalize the topology so web/desktop/mobile clients communicate only with Convex. Convex admits durable commands, dispatches private Flue turns through FLUE_URL, and stores the product-facing result before clients observe it. - Normalized relational schema: organizations, projects, conversations/turns/ messages/attachments, signals/sources/constraints, works/events/attachments. - Conversation turns queued in Convex; the agent runAction dispatches to Flue and persists assistant response, signals, and proposed work back into Convex. - Browser chat moved off the Flue transport: use-chat-agent now reads reactive Convex rows and sends through conversationMessages.send. - Fixed signed-in blank screen: gate all product queries on useConvexAuth (isAuthenticated && !isRefreshing) plus personal-org bootstrap. - Pinned react/react-dom to 19.2.8 via catalog to fix SSR dual-React crash. - Removed ~16.6k net lines: daemon, AgentOS/Orb, project issues/runs/artifacts, todos, browser Flue transport, mobile execution views, smoke scripts. Verified end-to-end on cheaptricks: connect repo, send actionable message, Flue turn completes, Signal + proposed Work created with exact provenance, persists across refresh.
41 lines
1.2 KiB
Plaintext
41 lines
1.2 KiB
Plaintext
# Shared application endpoints
|
|
CONVEX_DEPLOYMENT=dev:example-deployment
|
|
CONVEX_URL=https://example.convex.cloud
|
|
CONVEX_SITE_URL=https://example.convex.site
|
|
SITE_URL=http://localhost:5173
|
|
NATIVE_APP_URL=code://
|
|
|
|
# Browser and native public endpoints
|
|
VITE_AUTH_URL=http://localhost:5173
|
|
VITE_CONVEX_URL=https://example.convex.cloud
|
|
EXPO_PUBLIC_CONVEX_URL=https://example.convex.cloud
|
|
EXPO_PUBLIC_CONVEX_SITE_URL=https://example.convex.site
|
|
|
|
# Daemon identity and control plane
|
|
DAEMON_ID=local-macbook
|
|
DAEMON_NAME=Local MacBook
|
|
DAEMON_VERSION=0.0.0
|
|
DAEMON_HEARTBEAT_MS=15000
|
|
DAEMON_COMMAND_LEASE_MS=60000
|
|
# RIVET_ENDPOINT=http://localhost:6420
|
|
|
|
# Flue persistence adapter
|
|
FLUE_DB_TOKEN=replace-with-a-long-random-token
|
|
FLUE_URL=http://localhost:3583
|
|
|
|
# Agent model provider
|
|
AGENT_MODEL_PROVIDER=xiaomi
|
|
AGENT_MODEL_NAME=mimo-v2.5
|
|
AGENT_MODEL_API=openai-completions
|
|
AGENT_MODEL_BASE_URL=https://ai.example.com/v1
|
|
AGENT_MODEL_API_KEY=replace-with-provider-api-key
|
|
AGENT_MODEL_CONTEXT_WINDOW=1048576
|
|
AGENT_MODEL_MAX_TOKENS=131072
|
|
|
|
# Self-hosted git (Gitea) — canonical repository host and API token
|
|
GITEA_URL=https://git.openputer.com
|
|
GITEA_TOKEN=replace-with-a-gitea-personal-access-token
|
|
|
|
# Zopu lean server (standalone chat + work actor)
|
|
VITE_ZOPU_SERVER_URL=http://localhost:3590
|