AGI Agents
A streamlined interface for LangChain AI agent creation with multi-modal support (text and images).
Features
- Simple API: Easy-to-use interface for creating AI agents
- Multi-modal Support: Handle both text and image inputs seamlessly
- SSE Streaming: Built-in Server-Sent Events transport with middleware profiles (chat/tool/subagent)
- LangChain Integration: Built on top of the powerful LangChain framework
- Flexible Models: Support for OpenAI, Anthropic, Google GenAI, and other LLM providers
- Document Processing: PDF and image processing capabilities
- Context Management: Advanced context engineering utilities
- Async Support: Both synchronous and asynchronous execution
Installation
pip install agi-agents
Quick Start
from agi_agents import Agents
from langchain_openai import ChatOpenAI
# Initialize your model
llm = ChatOpenAI(model_name='gpt-4o-mini')
# Create a simple text processing chain
chain = Agents.chain_create(
model=llm,
text_prompt_template="Answer this question: {question}",
)
# Use the chain
response = Agents.chain_batch_generator(
chain,
{"question": "What is artificial intelligence?"}
)
print(response)
Multi-modal Example
# Create a chain that processes both text and images
chain = Agents.chain_create(
model=llm,
text_prompt_template="Describe this image: {description}",
image_prompt_template=True
)
# Process an image
base64_image = Agents.normalize_image_to_base64("path/to/image.jpg")
response = Agents.chain_batch_generator(
chain,
{
"description": "What do you see?",
"base64_image": base64_image,
"detail_parameter": "high"
}
)
SSE Streaming
from agi_agents.sse import StreamProfile
def my_agent():
yield "thinking..."
yield "final answer"
for chunk in StreamProfile.chat().wrap(my_agent(), model="gpt-4o-mini"):
# chunk is SSE-encoded bytes ready for text/event-stream responses
...
Built-in profiles: StreamProfile.chat(), StreamProfile.tool(), StreamProfile.subagent().
Main Classes
Agents
The core class for creating and managing AI agents:
chain_create(): Create LangChain processing chainschain_batch_generator(): Execute chains synchronouslychain_stream_generator(): Stream responses in real-timecontinue_chain_batch_generator(): Handle long conversations with continuation
Contexts
Utility class for context engineering:
create_context_layer(): Create context templatescompress_context(): Compress contexts to fit token limits
Document
Document processing utilities:
extract_text_from_pdf(): Extract text from PDFsconvert_pdf_to_images(): Convert PDFs to imagespdf_to_base64_images(): Convert PDFs to base64 for LLM processing
SSE (agi_agents.sse)
Server-Sent Events streaming transport:
StreamProfile: Declarative profiles for chat / tool / subagent streamsStreamPipeline/StreamMiddleware: Middleware pipeline (think, tool, audio)SSETransport/SSEEvent: Wire protocol encode/decode
Requirements
- Python 3.10+
- langchain / langchain-core 1.0+
- langchain-openai / langchain-anthropic / langchain-google-genai
- PyMuPDF
- Pillow
- pillow-heif
License
MIT License
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
Metadata
Release files for agi-agents 1.1.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agi_agents-1.1.0.tar.gz | 32.7 kB | Without provenance | Details for agi_agents-1.1.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agi_agents-1.1.0-py3-none-any.whl | Python 3 | none | any | Details for agi_agents-1.1.0-py3-none-any.whl |
Total release size: 65.3 kB
Release files / agi_agents-1.1.0.tar.gz
| Download URL | agi_agents-1.1.0.tar.gz |
|---|---|
| Size | 32.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
11c8e1fb81f875e7ea6a3338bc5551c3c1f8b3829dedc6637ae2f3b39863dfb6
|
|
BLAKE2b-256 checksum How to use checksums |
a7cfbea7a1fc489adb2b809baf6a6cdaeebb0b0b3c2e04472af15d8a903b9cf2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.13
|
Release files / agi_agents-1.1.0-py3-none-any.whl
| Download URL | agi_agents-1.1.0-py3-none-any.whl |
|---|---|
| Size | 32.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
a282257a15b37e1b54df6c3c3f62349ca69c883d9c5a59e41858ebd39fbf01a9
|
|
BLAKE2b-256 checksum How to use checksums |
64fdd1c69c8d5b11fa582f1a98c1ef01f9d9164905a40380eac95b19deafb0e5
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.12.13
|
Security for agi-agents 1.1.0
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.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.