Is SBproxy a reverse proxy, an AI gateway, or a model server?

All three, from one binary. It terminates TLS and handles HTTP/1.1, HTTP/2, WebSocket, and gRPC like nginx or Envoy. It routes LLM traffic to 200+ hosted models across 72 providers with guardrails, budgets, and MCP federation. And it serves open-weight models on your own GPUs by supervising vLLM or llama.cpp underneath. One config file drives all of it, and the AI features sit on a runtime that was already a production-grade proxy.

SBproxy is the open-source enterprise gateway for API, MCP, and AI model traffic. Outbound, it routes to 200+ hosted models across 72 providers through one OpenAI-compatible endpoint, picking models by prompt difficulty and rerouting on live cost-per-success, and serves open-weight models on your own GPUs. Inbound, it gates the crawlers, signed agents, and MCP clients that hit your APIs, with controls like Pay Per Crawl and Web Bot Auth. Both directions share one policy of virtual keys, budgets, and guardrails. Underneath sits a production-grade reverse proxy running auth, TLS, WAF, rate limiting, transforms, and caching. Single static binary, Apache License 2.0.

Can SBproxy serve models on my own GPUs?

Yes, on any GPU. A provider whose entire body is a serve: block boots to a first token: SBproxy resolves the weights, fits an engine and quantization to your card (bigger cards take FP8, smaller ones fall back to an int4 GGUF), runs vLLM or llama.cpp as a supervised subprocess, and routes to it. Run sbproxy doctor first and it reports what the host can serve. Every governance plane, meaning virtual keys, budgets, guardrails, and the usage ledger, applies to the local model unchanged, and a hosted provider placed after the local lane catches overflow under policy.

How is SBproxy different from vLLM or Ollama?

vLLM and Ollama serve a model. SBproxy stands up the environment around it: the governed OpenAI-compatible endpoint, virtual keys, budgets, guardrails on prompt and response, failover to hosted providers, and a verifiable usage ledger. It runs vLLM or llama.cpp underneath as supervised engines, so you keep the engine and gain the gateway.

How is SBproxy different from Cloudflare Pay Per Crawl?

Cloudflare runs at its own edge. SBproxy runs in your VPC, region, Kubernetes cluster, or air-gapped environment. Pricing is tier-aware, per route, per shape, and per agent rather than domain-wide. Inbound monetization, agent identity, content negotiation, and outbound AI governance live on one policy graph instead of separate Cloudflare products. Settlement supports both x402 and Stripe MPP, and you can act as your own Merchant of Record.

How is SBproxy different from other AI gateways like LiteLLM, Helicone, Portkey, or Kong AI Gateway?

Most AI gateways stop at routing calls to hosted providers. SBproxy also serves the weights: the same binary runs open models on your GPUs and applies the same keys, budgets, and guardrails to local and hosted traffic alike. It is also a real reverse proxy underneath, so auth, WAF, and rate limiting ride along instead of needing a second gateway. Helicone is excellent for LLM observability. LiteLLM is a Python library that grew a server, and sbproxy config import-litellm converts its config in an afternoon. Kong is a traditional API gateway with AI bolt-ons. None of them serve the model.

Where does SBproxy sit in my stack?

At the network hop between your apps or agents and the models they call, and between inbound AI traffic and your APIs. SDK gateways live inside one app's code, hosted edges govern from someone else's network, and agent platforms ask you to build on them; SBproxy is standalone infrastructure that deploys like nginx, works under any agent framework or platform, and fronts any hosted provider or your own GPUs. It terminates TLS with certificates it issues and renews from Let's Encrypt, so the governed endpoint can face the public internet behind virtual keys you mint and revoke at runtime.

How does SBproxy handle API keys and state across a fleet of instances?

SBproxy ships a dynamic key plane. Mint, rotate, and revoke virtual keys at runtime through the admin API with no reload, and a revoke takes effect on the next request. Keys are hashed at rest with HMAC-SHA256 and a server pepper, and upstream provider credentials are sealed with AES-256-GCM. Across a fleet of replicas, a built-in mesh keeps the keys, budgets, and per-key rate counters coherent using SWIM gossip and CRDT counters, so the cluster coordinates itself without an external Redis. LiteLLM and most AI gateways need a separate Redis to share that state across instances.

Where does SBproxy deploy?

Single binary on a server, Kubernetes operator on a fleet, or air-gapped install on a regulated network. Same policy graph in every shape. No phone-home. No SaaS edge. Bring your own Stripe, your own SSO, your own observability, and your own audit sink. Suitable for EU, India, Saudi Arabia, US public sector, defense, healthcare, and finance workloads with sovereignty requirements.

Is SBproxy open source?

SBproxy is open source under the Apache License 2.0. Free for any use, including production and commercial, with no field-of-use restriction. Read every line on GitHub. Verify every binary.

What is Pay Per Crawl and how does SBproxy support it?

Pay Per Crawl is the pattern of charging AI crawlers for access to content using HTTP 402 Payment Required. SBproxy ships the ai_crawl_control policy: AI crawlers that arrive without a valid payment token receive a 402 response with a JSON challenge. The crawler retries with the issued token and the proxy validates it through a pluggable ledger. Settlement supports x402 stablecoin and Stripe MPP. Tokens can be priced per route, per shape, and per agent.

What is Web Bot Auth and how does SBproxy verify AI agents?

Web Bot Auth is the IETF pattern for verifying signed AI crawlers and agents. SBproxy implements it through the bot_auth provider. Agents sign each request with an Ed25519 key under RFC 9421 HTTP Message Signatures and advertise their key id in the Signature-Input header. The gateway looks up the matching public key in its directory and verifies the signature. Agents that pass come through. Everything else gets 401.

Does SBproxy support MCP?

Yes. SBproxy ships an MCP gateway that federates internal tools under one identity model. Authentication supports OAuth2 with DPoP, PKCE, and DCR. Federated servers can be wrapped from any OpenAPI service, with per-server prefix and RBAC. The same policy graph that governs API and AI traffic also governs MCP tool invocations.

What does SBproxy replace?

A hosted model router (OpenRouter), an LLM proxy (LiteLLM, Portkey, Helicone), the hand-rolled shim in front of vLLM or Ollama, a traditional API gateway (Kong, nginx, Envoy), a guardrails service, a bot management product, and a payment rail for crawlers (TollBit, Skyfire). One runtime instead of six vendors stitched together, one policy plane instead of six control planes, one audit log instead of six disconnected feeds.

How many AI providers does SBproxy support?

SBproxy reaches 200+ models behind one OpenAI-compatible API through 72 native providers, including OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure OpenAI, Azure AI Foundry, Groq, Mistral, Cohere, Together AI, Fireworks, xAI Grok, OpenRouter, Cerebras, NVIDIA NIM, Databricks, Snowflake Cortex, IBM watsonx, and Oracle GenAI, plus the models it serves itself on your GPUs via vLLM or llama.cpp and external self-hosted runtimes like Ollama and TGI. Routing supports fallback chains, weighted, round-robin, latency-based, cost-optimized, and outcome-aware strategies.

Is SBproxy safe to run in front of production traffic?

SBproxy ships as a single static Rust binary. Releases are cut from tagged commits in a public repository on GitHub. Dependencies are pinned and checksummed in Cargo.lock. The runtime does not fetch code from a public package registry when it starts. The deployment shape matches nginx, HAProxy, and Envoy. Hot reload is supported without dropping connections.

CAP Tokens

The cap provider: verifying Crawler Authorization Protocol capability tokens (path globs, rate grants, agent binding) against an issuer JWKS

CAP (Crawler Authorization Protocol) is a JWT-based capability-token format for agent and crawler traffic. An agent presents a token in CAP-Token: <jwt> or Authorization: CAP <jwt>, and the token itself carries what the bearer may do: which paths, at what request rate, up to how many bytes a day. SBproxy ships the verifier side, the latency-critical half that runs on every request; tokens are minted by whichever issuer your deployment trusts, and the config names that issuer's keys.

This pairs naturally with the crawler-metering surfaces: ai-crawl-control.md challenges unpaid crawlers, and a CAP token is how a paid or contracted one proves its grant. The verified subject also feeds trust-tiers.md and access-log attribution.

Config

origins:
  "content.example.com":
    authentication:
      type: cap
      jwks_url: https://issuer.example.com/.well-known/cap/keys.json
      audience: content.example.com
      require_agent_binding: true

Field	Type	Default	Description
`jwks_url`	string	unset	Remote JWKS endpoint holding the issuer's public keys. One of `jwks_url` or `jwks_static` is required; `jwks_url` wins when both are set, so a deployment can rotate from static keys to a fetched set without a flag day.
`jwks_static`	object	unset	Inline JWKS document, for offline or pre-issued-token deployments with no issuer endpoint.
`jwks_refresh_secs`	int	`3600`	Remote JWKS cache interval, clamped to at least 30 seconds.
`audience`	string	request `Host`	Explicit audience the token's `aud` must equal. Unset means the token must name the host it is presented to.
`require_agent_binding`	bool	`false`	Require the token's `sub` to match the agent identity the resolver chain attached to the request. When set and no identity resolved, the request is refused: the binding fails closed.

What the verifier checks, in order

Signature against the JWKS (Ed25519 keys in the examples; anything the key set advertises).
Standard claims: exp not passed, iat not in the future, iss present, aud equal to the configured audience or the request Host, and cap_v == 1.
Subject binding: when the agent-class resolver chain has attached an agent_id to the request, the token's sub must match it. require_agent_binding: true makes a missing resolved identity a refusal rather than a pass.
Route grant: the request path must match the token's glob allowlist claim.
Rate grant: the token's rps claim is enforced in a per-subject token bucket, after everything above has passed, so an attacker cannot spend a subject's budget with bad tokens.

A verified token puts a principal on the request: the subject becomes the rate-limit and attribution identity, and the token id (jti) lands in the access log as cap_token_id, so per-token consumption is queryable.

What the caller sees

Outcome	Response
Verified	Request proceeds with the token's principal attached.
No token	`401` with `WWW-Authenticate: License`, the RSL 1.0 challenge, so a crawler discovers the scheme.
Token invalid (bad signature, expired, malformed, JWKS unreachable)	`401` with `WWW-Authenticate: License error="<code>"`, codes like `invalid_token`, `expired_token`, `directory_unavailable`. An unreachable JWKS refuses rather than admits: verification fails closed.
Token valid but not for this (wrong audience, subject mismatch, path outside `glob`)	`403` with codes like `invalid_audience`, `agent_mismatch`, `agent_binding_required`, `path_not_authorized`. The distinction is deliberate: `401` means retry with a better token, `403` means this token does not authorize this request.
Over the token's `rps` grant	`429` with `X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Reset`, and `Retry-After`.

Honest limits

The daily byte budget is not a hard cutoff. The token's bytes claim (max bytes per UTC day) is verified, surfaced on the principal, and attributable per token through the access log, but nothing in the current binary severs traffic when the budget is exhausted. Enforce it downstream off the attribution data, or treat it as a contractual term rather than a technical one.
Issuance is out of scope. SBproxy verifies; it does not mint tokens or run an issuer endpoint. The jwks_static shape exists precisely so a deployment can pre-issue tokens with its own tooling and hand the gateway only the public keys.
No licensing issuer, and no marketplace bridge. There is no RSL Open Licensing Protocol issuer here (the four-step flow that mints Ed25519-signed license tokens for AI crawlers), and no bridge to IAB's Content Authorization Marketplace Protocol. The gateway ships the verifying and challenging halves of crawler licensing: it can refuse an unlicensed crawler (ai-crawl-control.md), publish its terms (rsl.md), and verify a token somebody else minted. Minting and settling are the parts a licensing business runs, and they are not in this binary. If you see the issuer named anywhere as a thing to configure, that is a documentation bug rather than a hidden feature.

CAP Tokens

Config

What the verifier checks, in order

What the caller sees

Honest limits

See also