Pay for intelligence,
not inference.

USD of Rylvo-paid AI included each month — consumed by Workspace Architect, Agent Evolution, Prompt Optimization (variant generation + judging), Edge Case Engine, and Red Team (adversarial input generation). Bot execution during tests uses your own LLM key (BYOK).

Buy additional Architect Credits in $5/$25/$50/$100/$250 packs. Credits carry over; only included credits reset each month.

Build and edit bots by chatting with a natural-language AI that generates your full bot configuration.

Four reviewer LLM agents run on Rylvo's infrastructure — detect patterns, perf issues, output-quality, and validation gaps, then auto-generate guardrails. Draws from your Architect Credits.

OPRO, APE, DSPy Bootstrap, PromptBreeder — variant generation and LLM judging draw from Architect Credits. No monthly run cap; each run typically costs ~$0.20 in credits.

Detects failure patterns in production traces and auto-generates guardrails using Rylvo AI. Draws from your Architect Credits.

Adversarial input generation draws from Architect Credits; bot execution uses your own LLM key (BYOK). No monthly run cap; generation typically costs ~$0.15 in credits per run.

Unlimited on every plan — your bot chat uses YOUR LLM key (OpenAI, Anthropic, Google, OpenRouter, etc.). Rylvo routes the call server-side and logs a trace so you can debug, but adds zero markup on token costs. You pay your LLM provider directly.

Automated safety rules enforced at the verifier level.

Connect docs, files, URLs for retrieval-grounded bot responses.

Retrieval pipeline blueprints (chunking, embedding, rerank).

External tools and APIs your bots can call via MCP or webhooks.

Push every conversation to your own database or webhook — own your data end-to-end.

Storage events (conversation_created, message_appended, conversation_ended) included free per period.

USD per 1,000 events beyond the included pool. Deducted from your USD wallet; sync pauses if wallet hits $0.

Pro is a solo plan — one seat (the owner). Team adds collaboration with up to 25 seats plus custom roles & permissions.

Define your own roles (beyond Owner / Admin / Editor / Viewer) and assign granular capabilities.

Direct calls to POST /v1/respond (legacy decisioning API). Bot chat via UI is unlimited on all plans.

Per 1,000 additional backend requests beyond your monthly limit. Does NOT apply to bot chat (BYOK).

Base retention per plan. Extend with add-ons: +30 days ($5/mo), +90 days ($12/mo), +180 days ($20/mo), +365 days ($35/mo). Traces auto-delete after the retention window.

How long live oversight sessions and intervention logs are retained. Shares the same add-on as trace retention.

Distinct traces protected from auto-deletion via dataset membership. Preserved traces are exempt from the retention window.

Organize and label trace collections for evaluation, regression testing, or compliance.

Extend trace & Mission Control retention beyond your plan's base. +30 days ($5/mo), +90 days ($12/mo), +180 days ($20/mo), +365 days ($35/mo).

Re-run any past decision to debug failures without reproducing live traffic.

Stage heatmaps, action distribution charts, escalation trend analysis.

Real-time supervision of live conversations with intervention and audit-ready approvals.

Cron-based trace scans, reviews, report generation.

Slack, email, webhook, Discord, Teams.

Connect bots to external tools via the Model Context Protocol.

Inject your company's rules into the decisioning layer at the verifier level.

Dedicated inference pool with guaranteed lower latency.

Send messages to your audience across WhatsApp, Telegram, Email, SMS, Slack, Discord, Teams, and more. Supports templates, LLM content enrichment, and per-recipient tracking.

How long end-user / audience data is retained. Extend with add-ons: +30 days ($3/mo), +90 days ($7/mo), +180 days ($12/mo), +365 days ($22/mo).

Extend audience data retention beyond your plan's base. +30 days ($3/mo), +90 days ($7/mo), +180 days ($12/mo), +365 days ($22/mo).