Skip to main content

    Public Roadmap

    Full transparency into what we've shipped and what's coming next. We build in the open.

    Shipped

    12 features

    CORS-Ready API

    Live

    Call Behest directly from your browser. No backend proxy needed.

    Auth & Tenant Isolation

    Live

    Multi-tenant authentication with per-project API keys and JWT support.

    Three-Tier Rate Limiting

    Live

    Per-IP, per-project, and per-user rate limiting with zero code.

    PII Shield

    Live

    Automatic PII detection and protection before data reaches the LLM.

    Sentinel — Prompt Injection Defense

    Live

    Block jailbreak attempts and prompt injection attacks automatically.

    Conversation Memory

    Live

    Session-based conversation memory. Users pick up where they left off.

    System Prompt Management

    Live

    Configure your AI's personality and behavior per project.

    Token Budgets

    Live

    Automatic cost control with per-user and per-project daily limits.

    Kill Switches

    Live

    Instant emergency shutdown at global, tenant, or project level.

    Full Observability Stack

    Live

    Traces, metrics, and logs — correlated automatically.

    Usage Analytics & Spend Tracking

    Live

    Know your AI spend. Per-model, per-user analytics in real time.

    Self-Hosted Deployment

    Live

    Deploy in your cloud. Your data never leaves your infrastructure.

    Planned

    5 features

    Smart LLM Routing

    Coming Soon

    Route requests to the optimal model based on cost, latency, and capability.

    Semantic Cache

    Coming Soon

    Faster responses and lower costs for semantically similar queries.

    Built-in RAG

    Coming Soon

    Ground your AI in your documents with built-in retrieval.

    Usage Tiers & Token Economics

    Coming Soon

    Set up tiered pricing for your end users. Monetize your AI features.

    BYO LLM API Keys

    Coming Soon

    Bring your own OpenAI, Anthropic, Mistral, or any LLM API keys.

    Have a feature request?

    Let us know what you need →