Skip to main content

connector-oss — Python SDK

Tamper-proof memory, chain-of-custody, and OS-grade runtime for AI agents. Native Rust kernel via PyO3 — ~140 methods, zero Python overhead.

Install

pip install connector-oss

Prebuilt wheels available for:

  • Linux x86_64 / aarch64
  • macOS x86_64 / Apple Silicon (aarch64)
  • Windows x86_64

Quick Start

import os
from connector_oss import Connector

# 3 lines: agent with trusted memory
c = Connector("deepseek", "deepseek-chat", os.environ["DEEPSEEK_API_KEY"])
result = c.agent("bot", "You are helpful").run("Hello!", "user:alice")

print(result.text)         # LLM response
print(result.trust)        # 0-100 kernel-verified trust score
print(result.trust_grade)  # "A+" | "A" | "B" | "C" | "D" | "F"

From YAML Config

c = Connector.from_config("connector.yaml")

Memory & Knowledge

# Write tamper-proof memory (CID-addressed)
c.memory_write("pid:bot", "text", "Patient has fever", "user:nurse", "ns:er")

# Search memory
packets = c.search_namespace("ns:er", limit=10)

# Knowledge graph + RAG
c.knowledge_ingest("ns:er")
ctx = c.rag_retrieve("pid:bot", "ns:er", entities=["fever"], max_facts=10)

# Cognitive cycle
report = c.cognitive_cycle("pid:bot", "ns:er", "Diagnose patient")

Multi-Agent Pipeline

pipe = c.pipeline("support")
pipe.agent("triage", "Classify tickets")
pipe.agent("resolver", "Find answers")
pipe.route("triage -> resolver")
pipe.hipaa()

result = pipe.run("My account is locked", user="user:bob")
print(result.to_json())   # machine-parseable with provenance tags

Custom Folders (OS mkdir model)

c.create_agent_folder("nurse", "notes", "Patient notes")
c.folder_put("agent:nurse/notes", "p123", '{"bp": "140/90"}')
val = c.folder_get("agent:nurse/notes", "p123")

Observability

result.to_json()       # every field tagged with source (kernel/llm/derived/user)
result.to_otel()       # OTLP-compatible trace export
result.to_llm()        # LLM-friendly summary
result.provenance()    # {"kernel_verified": 5, "trust_percentage": 100.0}
result.is_verified()   # True = zero-fake guarantee

What Makes This Different

LangChain CrewAI connector-oss
Tamper-proof memory ✅ CID-addressed, kernel-verified
Audit trail ✅ Ed25519-signed, append-only
HIPAA/SOC2/GDPR ✅ Built-in compliance
Trust scoring ✅ Per-response, 0-100
Policy enforcement ✅ Non-bypassable 5-layer guard

Links

License

Apache-2.0

Release files for connector-agent-oss 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for connector-agent-oss 0.2.0
File Size Uploaded
connector_agent_oss-0.2.0.tar.gz 591.4 kB Details

Built distributions (wheels)

Table of built distributions (wheels) for connector-agent-oss 0.2.0
File
connector_agent_oss-0.2.0-cp313-cp313-manylinux_2_31_x86_64.whl CPython 3.13 CPython 3.13 Linux glibc 2.31+ x86-64 Details
connector_agent_oss-0.2.0-cp312-cp312-win_amd64.whl CPython 3.12 CPython 3.12 Windows x86-64 Details
connector_agent_oss-0.2.0-cp312-cp312-manylinux_2_34_x86_64.whl CPython 3.12 CPython 3.12 Linux glibc 2.34+ x86-64 Details
connector_agent_oss-0.2.0-cp312-cp312-macosx_11_0_arm64.whl CPython 3.12 CPython 3.12 macOS 11.0+ ARM64 Details
connector_agent_oss-0.2.0-cp312-cp312-macosx_10_12_x86_64.whl CPython 3.12 CPython 3.12 macOS 10.12+ x86-64 Details

Total release size: 21.8 MB

Release files / connector_agent_oss-0.2.0.tar.gz

Download URL connector_agent_oss-0.2.0.tar.gz
Size 591.4 kB
Tags Source
SHA-256 checksum
How to use checksums
aee763a19847c428c41d0df146c80f7e1fa5b0366a16d1bdce81f32ac049772b
BLAKE2b-256 checksum
How to use checksums
2b949a94b69eb2f0582d05327f819be75699c705a31694c44704d84b3336319a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.5

Release files / connector_agent_oss-0.2.0-cp313-cp313-manylinux_2_31_x86_64.whl

Download URL connector_agent_oss-0.2.0-cp313-cp313-manylinux_2_31_x86_64.whl
Size 5.9 MB
Tags CPython 3.13 Linux glibc 2.31+ x86-64
SHA-256 checksum
How to use checksums
9692bc1b062294b9c05cd249a3dcfcb0046dce80bf4558e792d4945f8ea19e8a
BLAKE2b-256 checksum
How to use checksums
0e2e69542177a07976ed39f1f8c23d736bb653c82dd8eb9aae9ddda4d9ba2d0e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.3

Release files / connector_agent_oss-0.2.0-cp312-cp312-win_amd64.whl

Download URL connector_agent_oss-0.2.0-cp312-cp312-win_amd64.whl
Size 4.6 MB
Tags CPython 3.12 Windows x86-64
SHA-256 checksum
How to use checksums
552ff5944e43c0ce6850155059adb491370ab5975bd10ee4e34151f4c77b0a1b
BLAKE2b-256 checksum
How to use checksums
7f0f54386fe7e7304f04571cad347836cbcb2779b64d41398567993be242a757
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.5

Release files / connector_agent_oss-0.2.0-cp312-cp312-manylinux_2_34_x86_64.whl

Download URL connector_agent_oss-0.2.0-cp312-cp312-manylinux_2_34_x86_64.whl
Size 3.8 MB
Tags CPython 3.12 Linux glibc 2.34+ x86-64
SHA-256 checksum
How to use checksums
c778b2d25aeba69a988f84b1a56ff4877ca560430f492272876ebce1e1ec0ef7
BLAKE2b-256 checksum
How to use checksums
ad13a9c3109614eeebd53d5645e1a7b7cfbbefe2d92b598d0e9b34290edd5e77
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.5

Release files / connector_agent_oss-0.2.0-cp312-cp312-macosx_11_0_arm64.whl

Download URL connector_agent_oss-0.2.0-cp312-cp312-macosx_11_0_arm64.whl
Size 3.4 MB
Tags CPython 3.12 macOS 11.0+ ARM64
SHA-256 checksum
How to use checksums
314f69857956c8e450a1532ed05b85158f15197930a505ac31bf9b9120f45801
BLAKE2b-256 checksum
How to use checksums
1c449f7bd0c465c6503bea124c1a966e5f7259e629f4ba383a63d1245144cf75
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.5

Release files / connector_agent_oss-0.2.0-cp312-cp312-macosx_10_12_x86_64.whl

Download URL connector_agent_oss-0.2.0-cp312-cp312-macosx_10_12_x86_64.whl
Size 3.5 MB
Tags CPython 3.12 macOS 10.12+ x86-64
SHA-256 checksum
How to use checksums
932e4900140e6a210e7519b7a32bd3d60306e3c662cda4658ba093c94417f29f
BLAKE2b-256 checksum
How to use checksums
4e59a560fb078ad3d80ce6e974ce0e922fcda138548fcac830e28e5ac0fd88e8
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via maturin/1.12.5

Release history Release notifications | RSS feed

This release

0.2.0 This release

6 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page