Simple bot to interact with Open Source LLM's running in Ollama using Telegram

Find a file

Alexey Skobkin 5280fe8a2e All checks were successful continuous-integration/drone/push Build is passing Details continuous-integration/drone/pr Build is passing Details Fixing user request representation in the LLM request context (#40 ).		2024-11-04 02:24:00 +03:00
bot	Fixing user request representation in the LLM request context (#40 ).	2024-11-04 02:24:00 +03:00
extractor	fix #15 slog usage. Also adding more logging.	2024-03-12 22:07:22 +03:00
llm	Context prompt format fix.	2024-11-04 02:10:57 +03:00
stats	#26 Adding in-memory chat history support. Removing inline queries. Refactoring stats and message processing a bit. Also changing LLM request context building a bit. Also adding alias for summarization and some other small changes.	2024-10-28 00:35:35 +03:00
.drone.yml	Also changing image tag for build step.	2024-03-10 05:57:59 +03:00
.gitignore	initial. Draft of SIMPLE LLM bot for Telegram chat.	2024-03-08 05:18:45 +03:00
Dockerfile	Model configuration and small prompt improvements (#24 )	2024-08-16 00:47:07 +00:00
go.mod	Refactoring structure from single file to several separated services. Adding new feature: "summarize" to generate bullet points for provided link.	2024-03-10 04:51:01 +03:00
go.sum	Refactoring structure from single file to several separated services. Adding new feature: "summarize" to generate bullet points for provided link.	2024-03-10 04:51:01 +03:00
LICENSE	LICENSE file.	2024-03-08 05:19:26 +03:00
main.go	Model configuration and small prompt improvements (#24 )	2024-08-16 00:47:07 +00:00
README.md	REAMDE.md fix (#25 )	2024-08-16 00:59:04 +00:00

README.md

telegram-ollama-reply-bot

Usage

Docker

docker run \
  -e OPENAI_API_TOKEN=123 \
  -e OPENAI_API_BASE_URL=http://ollama.localhost:11434/v1 \
  -e TELEGRAM_TOKEN=12345 \
  -e MODEL_TEXT_REQUEST=llama3.1:8b-instruct-q6_K
  -e MODEL_SUMMARIZE_REQUEST=mistral-nemo:12b-instruct-2407-q4_K_M
  skobkin/telegram-llm-bot