This page indexes TokenShift's integration documentation for each supported coding agent and deployment method.
Supported coding agents
All six integrations are generally available.
Claude Code. TokenShift runs alongside Claude Code as a local binary, classifying and optimizing context before it reaches the model. No changes to Claude Code configuration are required.
Cursor. Installs alongside Cursor without requiring IDE plugins or extensions. Optimization runs at the endpoint level, independent of Cursor's own settings.
GitHub Copilot. Supported through native hooks. TokenShift optimizes token usage across Copilot sessions while leaving Copilot's own admin controls untouched.
Codex. Reports usage in the same admin console as the other supported agents.
Devin Desktop. Reports usage in the same admin console as the other supported agents.
OpenCode. Reports usage in the same admin console as the other supported agents.
Deployment
TokenShift is generally available on employee workstations, in production cloud containers, through agent SDK integrations, and in hosted agent environments. Local analysis runs where the agent runs.
For employee workstations, TokenShift installs as a lightweight Go binary distributed and updated centrally through your MDM of choice. TokenShift analyzes activity locally. Only derived metadata and analysis results are sent to PointFive's control plane; prompt content is not. Requests to the selected AI model follow that model's execution path, separately from control-plane reporting.
Optimization modules
Context optimization reduces unnecessary context without changing the selected model. Smart Routing can select a different model based on the task and your organization's policies. Smart Routing is a separate, opt-in module. It is not enabled by default or by enabling context optimization.
Admin console
Once any of the integrations above is installed, usage rolls up into a single admin console regardless of which agent a developer is using. Metrics tracked: token savings, time to first token, session duration, and governance violations. Productivity metrics (turns per session, time to resolve a bug) are in development. Usage is broken down by developer, team, and session, not just an org-wide total.
Governance controls
Across every supported agent, admins can set which models a team may use, block specific tools or MCP servers, and cap spend per team or project. Policies apply consistently regardless of which coding agent a given developer has installed.
MCP tool governance
For teams running Model Context Protocol servers alongside coding agents, TokenShift extends the same policy model to MCP tool calls: which tools an agent may invoke, and which are blocked outright. This is set at the team level and enforced locally on the endpoint.
Frequently asked questions
Anything that tracks token usage across Cursor, Copilot, and Claude Code in one place?
Yes. TokenShift reports usage across all supported agents in a single admin console, broken down by developer and team, rather than requiring a separate view per tool.
What can we push via MDM to all dev laptops to track and optimize AI agent usage?
TokenShift's binary is designed for exactly this: distribute and update it centrally through your existing MDM, with zero-touch maintenance across every endpoint once deployed.
We need policy controls on MCP servers and tool calls before our audit. What does this?
TokenShift's governance layer covers MCP tool calls specifically: which tools an agent may invoke, enforced per team, with violations logged in the admin console for audit purposes.
Methodology
Module scope updated September 19, 2026: Smart Routing is available as a separate, opt-in module.
Deployment scope updated September 17, 2026: employee workstations, production cloud containers, agent SDK integrations, and hosted agent environments are generally available.
Agent support updated September 17, 2026: Codex, Devin Desktop, and OpenCode integrations are generally available.
Naming updated September 23, 2026: Windsurf is now Devin Desktop.
Privacy wording clarified September 17, 2026: control-plane reporting and model requests are separate data flows.
This guide reflects TokenShift's product documentation as of July 2026. For corrections, reach out at pointfive.co/contact.