Self-guided AI workflows for oil and gas data
Connect your data sources, public datasets, and MCP clients so you can ask better questions and get useful answers without a custom integration project.
Connect your Oil & Gas tools with AI
PatchOps bridges your oil & gas tools with AI through secure connectors, automating workflows.
Built now for people who want to work directly with the data
PatchOps starts with the self-guided user: someone who already knows the question, wants access to the right data, and needs an AI workflow that can execute safely without a services engagement.
Self-guided users
Available now
Individuals and small teams connect their own tools, then use an AI client to ask questions, pull records, compare assets, and run repeatable research without waiting on a services team.
1:1 assisted users
Later
A higher-touch path for teams that want help shaping workflows, onboarding data sources, or translating repeated analysis into a guided operating process.
B2C and service vendors
Future path
Consumer-facing products and vendors providing managed services can eventually use PatchOps as the connector layer behind their own customer workflows.
How PatchOps Works
Connect your stack
PatchOps links your data platforms and tools through secure connectors so AI agents can access everything in one place.
Guide the AI
Agents drill down layer by layer using MCP to pick the exact tool, field, and parameters required for the task.
Execute safely
PatchOps validates and runs the request with the right API details, delivering results instantly.
Context Window Management
Progressive Tool Discovery
Agents navigate tasks by purpose so they never guess which connector to call.
Eliminate Decision Paralysis
Only the next relevant action is exposed, preventing hallucinations and context waste.
Get all producing wells for an operator in Loving County
Routes to the right data source
Writes Code To Execute Datasource Call
Code Execution
Agent Returns Operator Production Data
Scalable Tool Integration
Handle thousands of tools without context overload so every request stays precise.
Why Patch Ops Is Better
Scalable Tool Integration
Handle thousands of tools with no context overload.
Progressive Discovery
Guides agents from intent to action, step-by-step.
Superior Accuracy
Proven to improve AI agents' accuracy on complex tasks.
Deep Feature Access
Access hundreds of features per app, not just basic ones.
Unify Your Toolset
Connect external MCP Server for unified access.
Reliable Multi-App Workflows
Enables complex automation across multiple apps.
Install once, then drive PatchOps from your terminal
Install @patchops/cli from npm, sign in with PatchOps OAuth, and use the same MCP-backed connector catalog from local shells, scripts, and automation workflows.
Direct URL for routing
Your PatchOps instance URL tells the CLI which MCP endpoint to call.
OAuth or PAT auth
Use PatchOps OAuth on the shared MCP endpoint or personal access tokens on direct instance URLs.
Revocable by design
Rotate or revoke CLI access without changing every instance URL you have shared internally.
CLI quick start
npm install -g @patchops/cli
patchops login
patchops tools listUse patchops invoke for direct automation once you know the connector and method you want, or move to direct instance plus PAT auth on the dedicated CLI page.
Build against PatchOps with first-party SDKs
The TypeScript and Python SDKs wrap the public OAuth-protected MCP surface so local apps, desktop tools, and services can handle discovery, PKCE, and JSON-RPC cleanly.
SDK quick start
import { PatchOpsMcpClient } from "@patchops/sdk"
const mcp = new PatchOpsMcpClient({ baseUrl, accessToken })
await mcp.initialize()
const tools = await mcp.listTools()