OpenAPI generator for Rust

Generate Rust clients and Axum servers from OpenAPI.

Q: What does this OpenAPI generator for Rust emit?

openapi-to-rust emits shared Serde models, selected async Reqwest methods, typed API errors, optional SSE streams, and selected Axum server traits and routes.

openapi-to-rust turns OpenAPI 3.0 and 3.1 documents into shared Serde models, selected Reqwest methods, typed API errors, SSE streams, and Axum traits.

CI generates all 54 specs on every pull request; OpenAI and Anthropic outputs also compile.

Try it in your browser Install the CLI

v0.8.0 · Rust 1.88+ · MIT licensed

openapi.yaml → src/generated/types.rs wasm

Message:
  type: object
  required: [id, role]
  properties:
    id:   { type: string }
    role: { $ref: "#/components/schemas/Role" }

↓ the real generator, in your browser — about 5 seconds

pub struct Message {
    pub id: String,
    pub role: Role,
}

Paste your own spec — nothing is uploaded. Open the playground

What CI verifies.

Every pull request runs the generator over 54 supported documents. OpenAI and Anthropic generated crates also run cargo check with exact dependencies. The scheduled and manual tier is configured to compile all 54. One Swagger 2.0 document is skipped.

This establishes parse, generation, and compile coverage—not behavior against a live API.

Corpus checks shell

# Every pull request
SPEC_COMPILE_PARSE_ONLY=1 scripts/spec-compile.sh
scripts/spec-compile.sh anthropic openai

# Scheduled and manual CI
scripts/spec-compile.sh

Every pull request54 generate · OpenAI + Anthropic cargo checkactive

Scheduled / manualAll 54 run cargo checkconfigured

Inspect the generated Rust.

This enum comes from the checked-in discriminator fixture. The repository also includes a generated Reqwest client and a compiled Axum implementation for OpenAI’s Responses API.

Model fixtureTagged Serde enum Client outputReqwest methods and typed errors Server exampleAxum traits, routes, and SSE

Generated types.rs rust

#[derive(Debug, Clone, Deserialize, Serialize)]
#[serde(tag = "type")]
pub enum ToolResultCaller {
    #[serde(rename = "direct")]
    DirectCaller(DirectCaller),
    #[serde(rename = "server")]
    ServerCaller(ServerCaller),
}

Generate only the operations you use.

Select an exact operationId, METHOD /path, or tag:name. Optional pruning keeps schemas reachable from the selected client and server operations.

Shared models: One types.rs for request and response shapes on both sides.
Reqwest client: Async methods, auth, typed responses and errors, retries, tracing, and SSE.
Axum server: Per-tag traits, response enums, parameter extraction, routing, and SSE responses.

Configure client generation Configure Axum generation

Commit generated code. Fail CI when it drifts.

--check exits nonzero when committed output no longer matches the OpenAPI document or configuration. Use --dry-run to inspect the generation plan without writing files.

Read the complete workflow

Project CI shell

openapi-to-rust generate openapi.yaml --check --quiet
cargo check

Before you adopt it.

openapi-to-rust is intentionally Rust-only. If you need many target languages, compare it with OpenAPI Generator’s separate Rust client and Rust Axum server generators.

Compare features and tradeoffs

What does this OpenAPI generator for Rust emit?

Shared Serde models, selected async Reqwest methods, typed API errors, optional SSE streams, and selected Axum traits and routes.

Which OpenAPI versions are supported?

OpenAPI 3.0 and 3.1 are supported. OpenAPI 3.2 parsing is experimental. Swagger 2.0 documents are not accepted.

Does generated code validate every schema constraint at runtime?

No. Minimums, maximums, and patterns are emitted as Rust documentation rather than runtime validators.

Start with your spec.

cargo install --locked openapi-to-rust

Generate Rust code View the source Open on crates.io