SkillCommunity★ 2.7k

SalesGPT

Updated Jul 7, 2026

SalesGPT by filip-michalsky is an open-source Python framework for context-aware AI sales agents. Tracks conversation stage (introduction → qualification → pitch → objection handling → close), searches a product catalog, generates Stripe payment links, sends emails, schedules Calendly meetings, and supports voice with under-1s latency.

View on GitHub

What it does

pip install salesgpt
Install the SalesGPT Python package.
python run.py --verbose True --config examples/example_agent_setup.json
Run a configured AI sales agent from the command line using your product catalog and persona config.
ProductSearch tool
Agent tool that searches the configured product catalog to answer prospect pricing questions accurately.
GeneratePaymentLink tool
Agent tool that autonomously generates a Stripe payment link when a prospect is ready to buy.

Files (1)

OVERVIEW.mdprimary · markdown · 1.5 KB

## What it does

SalesGPT is an open-source Python framework for building context-aware AI sales agents. It tracks which stage a sales conversation is in (Introduction, Qualification, Value Proposition, Needs Analysis, Solution Presentation, Objection Handling, Close) and adapts responses accordingly. Key capabilities: product catalog lookup via the ProductSearch tool, autonomous Stripe payment link generation to close orders, automated email sending, Calendly meeting scheduling, voice channel support with under-1-second round-trip latency, LangSmith tracing for debugging, and support for 50+ LLMs via LiteLLM. Configure it with your product catalog, company info, salesperson persona, and talk tracks. Run as a CLI script, Docker container, or backend API.

## Install

pip install salesgpt

Or for full setup with a local clone:

git clone https://github.com/filip-michalsky/SalesGPT.git && cd SalesGPT && make setup

Create a .env file with OPENAI_API_KEY (or your chosen LLM provider key). Run an example agent:

python run.py --verbose True --config examples/example_agent_setup.json

Docker: docker-compose up -d (then access at localhost:3000/chat).

## When to use

Use this when you want a programmable sales agent that tracks conversation stage and can autonomously search your product catalog and generate payment links — beyond just drafting email templates. For simple prompt-only use (no coding, no running Python), copy the system prompt from examples/ into a Claude Project instead.