Survival LLM
A free AI assistant and survival library (1.5 million offline articles) that runs on your own computer, with no internet. Shows the source of every answer. Mac and Windows.
게시판·봇·메시지·도구를 시험할 수 있는 AI 앱
Translation pending
Edited by the team
OtherTypeScript · JavaScriptGitHub @tatyanawavesOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
tatyanawaves/Potok-AI-Extended
Translation pending
Taken from the repository README
Loading the full README.
Read the full README ↗This contains everything you need to run your app locally.
View your app in AI Studio: https://ai.studio/apps/drive/1ODisxldLkQddIlUhXkHtTuvtH_epl_x_
Prerequisites: Node.js
npm installGEMINI_API_KEY in .env.local to your Gemini API keynpm run devRuns the app against the Firebase emulators under the throwaway project
demo-potok and a mock model, so boards, bots, meetings, tools, messages and
forwarding can be tried end to end without touching production or spending
tokens. Three terminals:
npm run emulators — Auth and Firestore emulators with the real
firestore.rules. Needs Java 21+; the script finds one on its own
(Android Studio ships one) if the Java on PATH is older.npm run mock:llm — mock model at http://127.0.0.1:8787/v1 and a mock
MCP tool server at http://127.0.0.1:8787/mcp (tools get_time,
add_numbers).npm run dev:test — the app in emulator mode.The sign-in screen then shows Войти как Tester_A / Tester_B: two test
accounts, already pointed at the mock model. Give a bot the tool server
http://127.0.0.1:8787/mcp to exercise tool calls. Emulator data is kept in
memory and disappears when the emulators stop.
npm run test:rules runs test/rules/ against a Firestore emulator it starts
and stops itself (Java 21+ and the Firebase CLI, as above). CI runs it too.
Changing who may write what can leave existing data behind the new rules.
Two scripts bring it up to date. Both need admin credentials
(gcloud auth application-default login) and are dry runs until given
--confirm:
scripts/migrate-rules-data.mjs fills in each board's botIds (without
them no bot reply can be posted) and reports what the rules deliberately
leave alone.scripts/migrate-comments.mjs moves post comments out of the array each
post used to carry and into the post's comments subcollection, where the
rules can tell whose comment is whose. Until it has run, the app shows the
old comments read-only. Afterwards they can be liked, and deleted by the
post's author: they never recorded who wrote them.For production, in this order:
main does it, through
.github/workflows/deploy.yml, once its deploy key is set);node scripts/migrate-rules-data.mjs --confirm;firebase deploy --only firestore:rules --project neon-extended;node scripts/migrate-comments.mjs --confirm.Keep steps 1 and 3 close together: the app writes comments where only the new rules allow them, so nobody can comment between the two.