MCP-native · no key required to start

Make the coding agent you already use dramatically smarter.

Helixor Code gives Claude Code and Codex a persistent, structural code graph — better than grep — and durable, stateful TODOs over MCP, shared across your team and surviving every session.

Request Access See how it works

Works with any LLM — or without one. No new IDE. No key required to start.

claude-code · helixor mcp

agent › where are auth tokens refreshed?
helixor › code.index.search — structural, not grep
auth/session.tsrefreshToken()def
api/client.tswithAuthRetry()usage
middleware/gate.tsrequireSession()bond
3 defs · 11 usages · 2 bonds · 142 tokens
todo ▸ stack "payments-refactor" · shared · 4 deep · survives session

      CODE INDEXES·STATEFUL TODOS·MCP-NATIVE·MULTI-LLM SDLC·TOKEN SAVINGS·SHARED MEMORY·
      CODE INDEXES·STATEFUL TODOS·MCP-NATIVE·MULTI-LLM SDLC·TOKEN SAVINGS·SHARED MEMORY·
    

Two ways to run it

Same product, from a single developer to a governed enterprise process.

Augment · no key

Drop it into your agent

Attach Helixor Code to Claude Code or Codex as an MCP server — better-than-grep lookup and stateful TODOs. The agent brings the model; you bring the tools and memory. Nothing to configure, no Helixor key.

Orchestrate · BYOK

Let it drive the SDLC

Let Helixor Code run a governed, multi-LLM SDLC on your own keys — Claude triages, a human approves, Codex implements, with evidence at every gate. Governed process, your key, your rates.

The flagship

One pipeline, the right model for each step.

Checkpointed, resumable, and artifact-backed at every step. This is what "AI that enforces your engineering standards" actually looks like.

See the full pipeline →

              intake
              →

              Claude triages
              →

              plan
              →

              human approval
              →

              Codex implements
              →

              tests
              →

              review
              →

            release notes
          

sdlc.safe_change_delivery.v1 — every step checkpointed, retryable (3×), resumable, and producing a required artifact.

The half nobody else ships

Both your code graph and your engineering memory — persistent and permissioned.

Most tools give you in-session context that evaporates, or a code graph locked to one machine. Helixor Code keeps living structural indexes and durable, depth-aware TODO stacks on the platform — so they survive across sessions, machines, and IDEs. Private to you, or shared with the whole team.

Features

Everything the agent was missing.

Living Code Indexes

Structural indexes over your repos: definitions, usages, bonds, motifs, quality, freshness. Better-than-grep lookup for agents and humans.

Stateful TODO Stacks

Durable, depth-aware stacks that survive across sessions, machines, and IDEs; shared with your team via web admin.

MCP-Native (no key)

Attaches to Claude Code / Codex as an MCP server; upgrades the agent you already run.

Code Quality Assessments

Architecture rules plus positive and negative motifs, scored, with actionable findings.

Token Savings

Structural context instead of re-reading the repo; fewer tokens, better answers.

Works in your IDE

VSCode/Cursor and JetBrains plugins; adopt a project in seconds.

Solo or Shared

Private developer workspace, or permissioned, org-wide system of record.

Multi-LLM SDLC Workflows

Checkpointed, gated pipelines that route different models to different steps and require a human at the gate — on your keys (Enterprise).

Cross-agent handoff

Shared workspace channels let Claude and Codex post, pin decisions, and hand off work auditably across sessions.

Any LLM, or none

Model-agnostic by design; the Helixor engine adds deterministic structure regardless of model.

Plans

Start free. Upgrade when you orchestrate.

Team

Free

Augment mode for individual developers and small teams. No key, no migration.

✓Living Code Indexes — structural, better than grep
✓Stateful TODO stacks over MCP
✓MCP-native for Claude Code & Codex, no key
✓Code quality assessments
✓VSCode/Cursor & JetBrains plugins
✓Token savings from structural context

No SDLC orchestration — augment only.

Request Access

Multi-LLM SDLC

Enterprise

BYOK

Everything in Team, plus governed orchestration on your own keys and shared team memory.

✓Everything in Team
✓Multi-LLM SDLC workflows (safe change delivery)
✓Human approval gates + artifact per step
✓Cross-agent workspace channels & handoff
✓Shared, permissioned team memory (web admin)
✓Governed orchestration on your keys (BYOK)

Each model call runs on your key, at your rates.

Talk to us

Build with Helixor Code

Make Claude Code and Codex smarter.

Request access for platform onboarding. Self-serve isn't open yet.