SalesGPT

Report Abuse

Basic Information

SalesGPT is an open source, context-aware AI Sales Agent designed to automate and manage sales conversations across channels including voice, email and various messaging platforms. The repository provides a Python-based agent implementation that understands sales conversation stages, leverages product and business knowledge to reduce hallucinations, and can execute actions such as generating Stripe payment links and scheduling meetings. It is intended for teams who want a deployable sales agent backend and example frontend, with options to run locally, via Docker, or integrate only the backend API. The project integrates with LangChain concepts, supports many LLM providers via LiteLLM, and includes tooling for tracing, human-in-the-loop supervision, and connecting to external data systems via Mindware. Installation options include pip or source, and examples and configuration files are provided to seed agents and run demos.

Links

App Details

Features
The repo implements stage-aware conversation management with predefined stages like introduction, qualification, needs analysis, presentation, objection handling and close. Agents can reference a product catalog and business knowledge to reduce hallucinations and call external tools for product search and payment link generation. Integrations include LiteLLM for using 50+ LLMs, LangChain patterns and LangSmith tracing, Mindware for connecting to arbitrary data systems, and planned security integration with PromptArmor. It supports synchronous and asynchronous LLM completions and streaming, low-latency voice pipelines with under 1s round-trip performance, automated email composition, Calendly meeting scheduling, Docker and backend-only deployment modes, example scripts, and developer-focused setup and test commands.
Use Cases
SalesGPT helps businesses automate repetitive sales workflows, qualify leads, answer product questions reliably using a controlled knowledge base, and actually close transactions by generating payment links. It scales sales interactions across voice and messaging channels while allowing human oversight where required. Developers and operators benefit from ready-made deployment options, backend APIs, tracing and test tooling to monitor and improve agent behavior, and integrations that let the agent access legacy or third-party data. The project provides example demos and a seeded agent workflow to accelerate building custom sales agents, reducing time to pilot and enabling production-ready capabilities like scheduling, email follow-ups, and low-latency voice interactions.

Please fill the required fields*