Skip to main content

DVARA 1.0.0-GA — Governance for Every LLM and MCP Call

· 7 min read
DVARA Team
DVARA Team

DVARA 1.0.0 is generally available today. We are launching an AI governance platform — the LLM Gateway and the MCP Proxy as components — for engineering teams running real production traffic and dealing with real auditors, real compliance officers, and real cost lines on a real budget.

This post covers why we built it, what is in 1.0.0-GA, and the angle that probably brought you here in the first place: DVARA Solo at $59 a month, flat, with no per-request overage.

Why DVARA exists​

If you have been routing LLM traffic through a Python proxy on a single VM, you already know the failure mode that pushed us to build this. Things work. Things keep working. Then production catches up.

Production catches up the day a compliance reviewer asks for proof that no PII left your perimeter, and the answer is a stack of unsigned text files. It catches up the day finance asks who spent the $14K on Anthropic last week and the answer is one shared API key. It catches up the day your first agent ships, hits a tool-call loop in the middle of the night, and burns through a budget cap that does not exist yet.

The existing options solve part of the problem. LiteLLM is excellent for prototyping — a single binary that hides the differences between OpenAI, Anthropic, and a long tail of providers. We respect what that project did for the ecosystem; this post is not an attempt to argue otherwise. The constraint is that LiteLLM was built for a developer routing their own calls, not for a platform team governing many teams at once. Multi-tenant cost attribution, signed and chained audit trails, Policy-as-Code that a compliance reviewer can read, and tool-call governance for the agents your company is about to ship — those problems land somewhere else.

Portkey is a managed SaaS that adds observability and a unified API. It is more polished than rolling your own — and it is also priced per call. Above the included quota, every request is a metered line item. For a team trying to forecast cloud spend twelve months out, that is a planning problem before it is a billing problem. SaaS shape, SaaS uncertainty.

DVARA is what you deploy when the workload is past prototyping and the people asking questions about it are no longer all engineers. Drop-in OpenAI compatibility is still there — we kept the easiest part — but the centre of the product is governance. Every call is policy-evaluated, PII-scanned, audited, and cost-attributed before it leaves your perimeter, on call one. Routing, failover, structured outputs, and semantic cache exist in the same product, but they exist to serve the governance story, not the other way round.

What is in 1.0.0-GA​

The Day-1 surface of DVARA is the managed shared cluster — the same code you would self-host, run by us, multi-tenant, with real isolation between tenants. Four pricing tiers cover from "evaluating" to "running an agent platform":

  • Trial — free for 30 days, 10 million tokens per month, every governance capability unlocked, no credit card. One per company domain, not renewable. After 30 days, continue on Solo.
  • Solo — $59 a month, flat. 10 million tokens included per month, all 14 providers, BYOK, signed audit, PII detection, basic Policy-as-Code, basic cost attribution, DVARA MCP Proxy included. No per-request overage. The right tier for a solo developer or a 2-3 person AI startup shipping a vertical product.
  • Starter — $299 a month. 25 million tokens, plus advanced Policy DSL, plus SIEM export, plus budget caps with auto-downgrade, plus the unified LLM + MCP session timeline. The right tier for a 3-7 engineer team running production AI.
  • Growth — $499 a month. 50 million tokens, plus 1-year audit retention, plus priority email + Slack support, plus quarterly compliance review. The right tier for teams running production AI at scale.

The platform supports OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure OpenAI, Ollama, Mistral, Cohere, Groq, Qwen, DeepSeek, Moonshot Kimi, ChatGLM, and xAI Grok on Day 1, all through a drop-in OpenAI-compatible interface. Capability-aware route filtering means a request asking for json_schema does not get routed to a provider that cannot natively support it. Failover, structured outputs, streaming, and a per-route circuit breaker are along for the ride.

Governance is the centre. Every event in the audit trail is cryptographically signed and chained at write time, so any tampering is detectable end-to-end. The Policy-as-Code engine takes a YAML DSL that a compliance reviewer can read, supports dry-run against historical traffic, and can be promoted from SHADOW to ACTIVE without restart. PII detection scans inputs and outputs and supports BLOCK / REDACT / LOG actions per tenant. Cost attribution lives at the request level, not the daily aggregate, so the spend for one tenant on one model on one day is a query, not a forensic exercise. SOC 2, HIPAA, and GDPR evidence packages are generated on demand from the audit trail.

For the audiences this Day-1 launch is not for — Team, Business, and Enterprise — productization lands in Q3 2026, with both managed-dedicated (we run a cluster in your region) and self-host (license key, deploy in your VPC) deployment options. Until those tiers ship, we set those customers up white-glove. If you want founder-led help getting DVARA to production today, our Pilot offer is a 4-6 week engagement; talk to us about it.

Solo at $59 vs Portkey​

The reason you might be reading this post is that you are running on Portkey today, watching your call volume climb, and feeling the line items climb with it. Solo at $59 a month, flat, is built for you.

Same volume as Portkey's $49 base tier, plus 10 million tokens included, with no per-request charge above the cap until you choose to upgrade. You stay on the same line item until your traffic outgrows Solo and you move to Starter. No surprise invoice the month a customer's chatbot suddenly takes off.

Same drop-in OpenAI compatibility, so the migration is two lines of code. Change the base_url in your OpenAI SDK from https://api.portkey.ai/v1 to https://api.dvarahq.com/v1, paste your provider keys in the DVARA Flightdeck, and you are done. Existing application code keeps running. Streaming, structured outputs, tool calls, all of it.

The bigger differentiator is governance shape. Where Portkey's product surface is observability for engineers, DVARA's is governance for compliance, FinOps, and platform teams. Signed and chained audit. Per-tenant Policy-as-Code. PII detection that actually does something. And the DVARA MCP Proxy — which has no equivalent in Portkey today — included on every paid tier from Solo up, for teams whose roadmap includes agents.

We wrote a longer migration walkthrough at /migrate-from-portkey. Same shape from LiteLLM Cloud at /migrate-from-litellm-cloud.

If you want to see DVARA on your own traffic, start a 30-day trial — no credit card, all features unlocked. If you have a production workload and want founder-led help getting DVARA into your VPC, email us about a pilot — $30–50K for 4-6 weeks, includes a year of the Growth tier and a published case study. Six pilots a year, capped at one active pilot at a time.

The product page lives at /pricing. The launch is live now.