NeuralBridge: The Universal Enterprise Middleware for Agentic AI
The missing infrastructure layer for agentic AI. Proudly developed by Grimaldi Engineering.
NeuralBridge is an open-source, standardized adapter platform that lets ANY AI agent (OpenClaw, AutoGPT, LangChain, Claude, ChatGPT) securely connect to ANY data source, API, or enterprise system with simple YAML configuration.
The Enterprise Problem: Agents Can Think, But They Can't Act
The agentic AI ecosystem is exploding. Projects like OpenClaw prove that agents can reason, plan, and learn. But a critical gap prevents their adoption in the enterprise: how do they act in the real world—safely, securely, and at scale?
Enterprises are eager to deploy AI agents to automate complex workflows, but are blocked by fundamental challenges:
- Security & Compliance: How do you grant an LLM access to a production database without risking data exfiltration or violating GDPR?
- Integration Complexity: How do you connect an agent to a dozen different systems (Salesforce, SAP, Snowflake, etc.) without writing thousands of lines of brittle, custom code?
- Lack of Standardization: How do you manage a fleet of agents from different vendors, all speaking different protocols?
NeuralBridge is the answer. It provides the secure, standardized infrastructure layer that enterprises need to deploy agentic AI with confidence.
Why NeuralBridge? The Value Proposition
| Feature | Description |
|---|---|
| Universal Adapter Framework | With 22+ pre-built adapters, NeuralBridge is the "Stripe for AI agents." Connect to any database, API, messaging platform, or enterprise system in minutes using simple YAML. |
| Built-in EU CRA Compliance | NeuralBridge is designed for the future of AI regulation. It comes ready for the EU Cyber Resilience Act (deadline Sept 2026) with immutable audit logs, SBOM generation, and vulnerability reporting. |
| Zero-Trust Security Engine | Every connection is protected by role-based access control (RBAC), credential encryption (via Vault), rate limiting, and a secure sandboxing engine that isolates every agent action. |
| No-Code Configuration | Business users can connect agents to systems via a beautiful React dashboard. No coding required to create, secure, and deploy a new connection. |
| Agent-Agnostic Gateway | A standardized entry point for any agent using the Model Context Protocol (MCP), with native support for the OpenClaw plugin ecosystem. |
Quick Start
1. Docker Compose (Recommended)
docker compose up -d
Visit the dashboard at http://localhost:3000 and the API docs at http://localhost:8000/docs.
2. pip
pip install neuralbridge-middleware
neuralbridge serve
Architecture Overview
See the full Architecture Documentation for a deep dive into the components.
Feature Deep Dive
Universal Adapter Framework
Connect to anything with our 22+ adapters:
- Databases: PostgreSQL, MySQL, MongoDB, Snowflake, BigQuery
- APIs: REST, GraphQL, SOAP, OData
- Messaging: Slack, Teams, Discord, Telegram, Email
- Productivity: Gmail, Notion
- ERP / CRM: Salesforce, SAP
- Cloud: AWS S3, Azure Blob, Google Cloud Storage
EU CRA Compliance Engine
NeuralBridge is designed for the future of AI regulation.
- Immutable Audit Logs: Hash-chain-verified logs of every agent action.
- SBOM Generation: CycloneDX-compliant Software Bill of Materials.
- Vulnerability Reporting: Generate CRA Article 14 reports on demand.
React Dashboard
- No-Code Connection Wizard: Intuitive UI for non-technical users.
- Compliance Monitoring: Real-time CRA and GDPR readiness status.
- Cost Optimization: Token cost analytics and caching recommendations.
Example: Connecting an Agent to Salesforce
Create a salesforce.yaml file:
adapters:
salesforce:
type: salesforce
auth:
type: oauth2
client_id: ${SALESFORCE_CLIENT_ID}
client_secret: ${SALESFORCE_CLIENT_SECRET}
instance_url: https://yourorg.my.salesforce.com
permissions:
- query: "SELECT Id, Name FROM Account"
Run the connection:
neuralbridge connect --config salesforce.yaml
Now, any agent can use the salesforce_query tool through the NeuralBridge MCP gateway.
Contributing
We welcome contributions from the community! Please see our Contributing Guide to get started.
Security
For security vulnerabilities, please see our Security Policy.
Roadmap
See our Roadmap for planned features.
License
NeuralBridge is licensed under the MIT License.
Metadata
Release files for neuralbridge-middleware 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| neuralbridge_middleware-0.1.1.tar.gz | 113.4 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| neuralbridge_middleware-0.1.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 258.7 kB
Release files / neuralbridge_middleware-0.1.1.tar.gz
| Download URL | neuralbridge_middleware-0.1.1.tar.gz |
|---|---|
| Size | 113.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a9e1b5a64ee8a7be8031dd66311da0e7d5f7c4a8dd4e4f697fd0d50185a51432
|
|
BLAKE2b-256 checksum How to use checksums |
dd6e250c10b3f51a4e3b1838f87b150ff205ba11181bd74ec6bc8c46b94f47f8
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
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 Mar 7, 2026.
Transparency logRelease files / neuralbridge_middleware-0.1.1-py3-none-any.whl
| Download URL | neuralbridge_middleware-0.1.1-py3-none-any.whl |
|---|---|
| Size | 145.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
d56311deeea809beff8bcbb1ddd8bacb4e9795804983add16e4c1013080d9a1d
|
|
BLAKE2b-256 checksum How to use checksums |
992fa957de4c66b203cacc721c84b876e1a200d51320694f77b20d030a20052d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
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 Mar 7, 2026.
Transparency log