Elyra Team · Early access June 2026

Your team uses AI agents. You have no idea what they cost, what they change, or what they know.

Elyra Team gives your organization cost control, shared context, and safety guardrails for AI-assisted development. The CLI stays free and open source. Team solves the coordination problem.

Join the waitlist How it works

01 · The problem

It works fine for one person. It breaks for a team.

Five developers using AI agents independently means five separate API keys, five separate billing accounts, no visibility into total spend, no shared context about the codebase, and no way to prevent an agent from deleting a production migration file at 2 AM.

02 · What Team adds

Three things Team solves

01

Managed Proxy

One bill, full visibility. Developers stop managing individual API keys. Team admins register provider credentials once. Every Elyra session routes through the proxy automatically.

What developers see
Nothing changes. They set one environment variable and Elyra works exactly as before. The proxy is invisible.
What admins see
A Filament dashboard showing every session, every dollar, every model choice. Export to CSV for finance. Set alerts for budget thresholds.
What you get
  • Single API key per team, distributed via ELYRA_PRO_KEY
  • Per-developer and per-project budget caps
  • Automatic provider failover (Anthropic down? Route to OpenAI)
  • Real-time cost dashboard: who spent what, on which project, with which model
  • Monthly trend graphs and alerts when spend approaches limits
02

Shared Memory

The agent remembers what your team knows. When a senior developer teaches the agent that your project uses the repository pattern, that knowledge should be available to every developer on the team — not locked in one person's ~/.elyra/ directory.

How it works
  • · Project memory lives in .elyra/memory/ as markdown files, versioned in git
  • · The agent writes summaries after each session: architecture decisions, patterns learned, gotchas discovered
  • · Every developer pulls the latest memory with their normal git workflow
  • · The agent reads it on session start and knows what the team knows
Team layer adds
  • Real-time sync without waiting for git push
  • Access control: who can write to memory, who can only read
  • Dashboard showing what the agent has learned, with review/approval workflow
  • Conflict resolution when two developers' sessions produce contradicting knowledge
Example

Junior developer starts a session. The agent already knows: "This project uses UUID primary keys, not auto-increment. Repository pattern for all database access. Tests go in tests/Feature/ and use Pest, not PHPUnit." Nobody had to explain it. The memory was there from the senior developer's sessions last week.

03

Guardrails

Rules the agent can't break. Configurable rules that prevent the agent from performing dangerous operations, even if the developer asks it to.

How it works
  • · Rules defined in .elyra/guardrails.yml or via the dashboard
  • · Enforced as pre-write hooks in the tool execution pipeline
  • · The agent sees the rule, explains why it can't proceed, and suggests an alternative
  • · Admins get notified when a guardrail triggers
Why this matters
The agent is powerful. That's the point. But power without boundaries is a liability in a regulated or mission-critical codebase. Guardrails let you give your team the full power of AI-assisted development with the safety net your CTO requires.
Examples
  • Never delete migration files that have been committed
  • Never modify auth middleware without manual confirmation
  • Never write to production environment files
  • Never remove test files
  • Always run tests after editing more than 3 files

03 · Architecture

How it works

One environment variable. No code changes. No workflow disruption.

request flow 
Developer's terminal
    |
    v
Elyra CLI  (free, open source)
    |
    | ELYRA_PRO_KEY
    v
Elyra Team proxy  (your infrastructure or ours)
    |
    |-- Authenticates developer
    |-- Checks budget
    |-- Selects provider key
    |-- Streams LLM response
    |-- Logs usage
    |
    v
LLM Provider  (Anthropic, OpenAI, Google, etc.)

04 · What stays free

The CLI stays free. Forever.

Everything the individual developer uses today stays free and open source.

Feature Free Team
CLI + all extensions
30+ LLM providers
Local sessions and memory
Themes, blueprints, snippets, pins
Smart routing and compaction
Managed proxy with cost dashboard —
Per-developer budget caps —
Shared team memory with sync —
Guardrails and safety rules —
Audit log (who changed what, when) —

05 · Pricing

Simple, predictable pricing

$25
per developer, per month

You bring your own provider API keys (Anthropic, OpenAI, etc.). We don't mark up token costs. You pay providers directly at their published rates.

Team pricing covers the proxy infrastructure, dashboard, shared memory sync, guardrails engine, and audit logging.

No per-token fees
No usage-based pricing
No surprises

06 · Early access

Join the waitlist

We're building Elyra Team for early access in June 2026. Leave your email and we'll reach out when it's ready.

Email
Team size What's your biggest pain point with AI agents in a team? Optional — but it really helps us prioritize.

We'll only use your email to contact you about Elyra Team. No marketing spam.

Elyra is free and open source. Team is for organizations that need cost control, shared context, and safety guardrails. The CLI will never require a Team subscription.

elyracode.com · GitHub · Docs