🗂️vCon Admin

A Streamlit toolkit for conserver developers, testers and operators to move vCons between stores, inspect them and run QA analysis.

Repo: vcon-dev/vcon-admin

vcon-admin is a Streamlit application for people who build, test and run a conserver. It uses MongoDB as its own data store and runs in Docker.

What it does

• Imports and exports vCons between Redis (the conserver's native store), S3, JSONL, JSON and MongoDB.

• Shows the status and configuration of the local system and its Docker containers, with live Docker logs.

• Uploads vCons to Elasticsearch, and to Milvus or an OpenAI vector store, for QA testing and semantic search.

• Inspects one vCon by UUID: summary, analysis, dialog, parties and attachments, with a JSON download.

• Runs a workbench that sends configured OpenAI prompts (full vCon, summary or transcript) over a chosen set of vCons, and a ChatGPT section for file upload and assistant chat.

• Visualizes vCon data and embeddings.

Install

The README describes a Docker install tested on Ubuntu 24.04:

1. Clone the repo and create .streamlit/secrets.toml.

2. Run docker network create conserver, then docker compose up -d.

3. Reset the Elasticsearch password with docker exec -it elasticsearch /usr/share/elasticsearch/bin/elasticsearch-reset-password -u elastic and put it in secrets.toml.

4. Open http://localhost:8501.

secrets.toml holds sections for AWS (S3 import and export), MongoDB, OpenAI, Elasticsearch and the conserver API URL and token. For Milvus, start docker-compose-milvus.yml as well. To run without Docker, use poetry install then poetry run streamlit run admin.py.

Do not expose port 8501 on a public host without a firewall rule.

When not to use it

• For LLM-driven queries, use the vCon MCP Server.

• For pipelines (transcribe, analyze, store), use the conserver.

See also

• Conserver

• Conserver Storage