🧠 IAToolkit — Open-Source Framework for Real-World AI Assistants
Build private, production-grade AI assistants that run entirely inside your environment and speak the language of your business.
IAToolkit is not a demo wrapper or a prompt playground — it is a full architecture for implementing intelligent systems that combine LLMs, SQL data, internal documents, tools, workflows, and multi-tenant business logic.
🚀 Why IAToolkit?
Modern AI development is fragmented: LangChain handles chains, LlamaIndex handles documents, your backend handles SQL, your frontend handles chats, and your devs glue everything together.
IAToolkit brings all of this into one unified, production-ready framework.
It focuses on:
- real-world data (SQL + documents)
- real workflows (LLM tools + python services)
- real multi-tenant architecture (1 company → many companies)
- real constraints (security, reproducibility, governance)
- real deployment (your servers, your infrastructure)
IAToolkit lets you build the assistant that your organization needs — not a generic chatbot.
🧩 Architecture in a Nutshell
IAToolkit is a structured, layered framework, with a clear separation of concerns.
✔ Interfaces
Chat UI, REST API, auth, sessions, JSON/HTML responses.
✔ Intelligence Layer
Core logic: prompt rendering, SQL orchestration, RAG, LLM tool dispatching.
✔ Execution Layer
Python services that implement real workflows: querying data, generating reports, retrieving documents, executing business logic.
✔ Data Access
A clean repository pattern using SQLAlchemy.
✔ Company Modules
Each company has:
- its own
company.yaml - its own prompts
- its own tools
- its own services
- its own vector store & SQL context
This modularity allows true multi-tenancy.
🔌 Connect to Anything
IAToolkit integrates naturally with:
- SQL databases (PostgreSQL, MySQL, SQL Server, etc.)
- Document retrieval (PDF, text, embeddings)
- External APIs
- Internal microservices
- Custom Python tools
It also includes a production-grade RAG pipeline, combining:
- embeddings
- chunking
- hybrid search
- SQL queries + document retrieval
- tool execution
Everything orchestrated through the Intelligence Layer.
🏢 Multi-Tenant Architecture
A single installation of IAToolkit can power assistants for multiple companies, departments, or customers.
companies/
company_a
company_b
company_c
Each Company is fully isolated:
- prompts
- tools
- credentials
- documents
- SQL contexts
- business rules
This makes IAToolkit ideal for SaaS products, agencies, consultancies, and organizations with multiple business units.
🆓 Community Edition vs Enterprise Edition
IAToolkit follows a modern open-core model:
🟦 Community Edition (MIT License)
- Full Open-Source Core
- SQL + Basic RAG
- One Company
- Custom Python tools
- Self-managed deployment
Perfect for developers, small teams, single-business use cases, and experimentation.
🟥 Enterprise Edition (Commercial License)
- Unlimited Companies (multi-tenant)
- Payment services integration
- Enterprise Agent Workflows
- SSO integration
- Priority support & continuous updates
- Activation via License Key
👉 Licensing information:
🧩 Who Is IAToolkit For?
- Companies building internal “ChatGPT for the business”
- SaaS products adding AI assistants for multiple customers
- AI teams that need reproducible prompts and controlled tools
- Developers who want real workflows, not toy demos
- Organizations requiring privacy, security, and self-hosting
- Teams working with SQL-heavy business logic
- Consultancies deploying AI for multiple clients
⭐ Key Differentiators
- prioritizes architecture-first design, not chains or wrappers
- supports multi-company out of the box
- integrates SQL, RAG, and tools into a single intelligence layer
- keeps business logic isolated inside Company modules
- runs entirely on your own infrastructure
- ships with a full web chat, and API.
- is built for production, not prototypes
📚 Documentation
- 🚀 Quickstart – Set up your environment and run the project
- ☁️ Deployment Guide – Production deployment instructions
- 🏗️ Companies & Components – how Company modules work
- 🧠 Programming Guide – services, intelligence layer, dispatching
- 🗃️ Database Guide – internal schema overview
- 🌱 Foundation Article – the “Why” behind the architecture
- 📘 Mini-Project (3 months) – how to deploy a corporate AI assistant
🤝 Contributing
IAToolkit is open-source and community-friendly.
PRs, issues, ideas, and feedback are always welcome.
⭐ Support the Project
If you find IAToolkit useful, please star the GitHub repo — it helps visibility and adoption.
1 maintainer
Release files for iatoolkit 2.51.2
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| iatoolkit-2.51.2.tar.gz | 701.3 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| iatoolkit-2.51.2-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 1.5 MB
Release files / iatoolkit-2.51.2.tar.gz
| Download URL | iatoolkit-2.51.2.tar.gz |
|---|---|
| Size | 701.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3c45e7a5e101cd15f5da24e4a09ddb08b02a0a82446d11b054455102ec3c6a1a
|
|
BLAKE2b-256 checksum How to use checksums |
777df3bfe601a41ef50c6b30fca7c7eb17053df695e12ee1ec6a87782319d07f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 25, 2026.
Transparency logRelease files / iatoolkit-2.51.2-py3-none-any.whl
| Download URL | iatoolkit-2.51.2-py3-none-any.whl |
|---|---|
| Size | 805.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
89e45bdc07dca6bb38958199e6455761b93b5d34a7cb435281d329c7ba6126ef
|
|
BLAKE2b-256 checksum How to use checksums |
1479bb234a8315adebeb76bc2ee203fbca6f32517ff0d0e80706fc9464f5fda7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 25, 2026.
Transparency log
