The control plane for coding agents
Oversized prompts. Broken boundaries. Endless retries. Give your coding agent five guardrails that work inside the editor you already love.
YOUR EDITOR › HEXUM
Every failure has a guardrail
Pick a problem. See what changes. Each product handles a different moment in your workflow.
01 / 05 · Efficient
Entire modules, repeated helpers, irrelevant context.
Task-shaped context. Expand only when the agent needs more.
hexum_shrink → relevant context hexum_expand → inspect a detail hexum_splice → edit the real source3 MCP tools · your existing editor
Try it on your own files
Try the sample repo or choose source files from your project. See how much context Hexum trims before connecting your agent.
Drop source from a real folder. Hexum folds what the model does not need this turn.
Up to 8 files, 96KiB. Nothing is stored. This playground is not the MCP.
A familiar workflow
Connect Hexum once in your editor’s MCP settings. Start on the free starter window, then choose a plan if you need more.
Set up your connectionNew to MCP? Follow the setup guide
{
"mcpServers": {
"hexum": {
"serverUrl": "https://hexum.dev/mcp",
"headers": { "Authorization": "Bearer hexum_live_…" }
}
}
}Code stays yours
No payload tables. No stored diffs.
Built around MCP
Use it with the editor you already have.
Start without a card
Personal keys include a monthly starter window.
Last updated 12 September 2026
Hexum is a developer ecosystem for coding agents in Cursor, Claude Code, and the rest. One API key. Five products and twelve tools today: spend less on each prompt, block bad imports, review pull requests, stop the agent looping, and only keep building when checks pass. We do not keep your code.
No. You keep Cursor, Claude Code, Antigravity, Codex, or Grok. Hexum is not an editor and not a model. You paste a key into the agent you already use.
Your agent stops dumping the whole repo into the prompt. Hexum sends this task, not every file. On the published card, MBPP context dropped 85.2% and the combined 24-task cut is 87.6%, with 24/24 tests still passing. Tokens are ceil(utf8_bytes/4), not a provider invoice. See /benchmarks.
No. It reads a rules file in your repo and reports forbidden imports. It does not call a model, so it does not guess.
You pay for what your agent sent Hexum. Usage €0.30 per million tokens. Teams €0.35 per million. No per-person fee. No monthly cap. See /pricing.
How much smaller the reply was than what the agent sent, when that is a cut. Your invoice is still based on what the agent sent Hexum, not on the savings.
No. We do not keep the code your agent sends. We do not train on it. Policy: /zdr.
Create a key at /app, paste the one-sentence prompt into Cursor, and let the agent write the MCP config. Starter works until two million tokens this month. Details: /cursor.
A file in the repo (usually architecture.hexum) that groups folders into layers and says which layer cannot import another. Guide: /architecture-hexum.
AUTH means the key is missing, wrong, or revoked. BILLING means that key’s plan is off, or the starter window is used. Turn Usage billing on at /app, or make a key on the plan that is on.
Hexum is an ecosystem: five products and twelve tools on one key. Efficient: don’t paste the whole repo. Arch: block a bad import. Review: architecture review on GitHub PRs. Control: when the agent loops, write the rule. Loop: only keep building if checks and tests pass.
Your tools. Your editor. More control.
Connect Hexum to the workflow your team already knows.
Get a key