# Sphynx

> Evals for coding agents

- [Introduction](https://docs.sphynx.sh/introduction.md): What is Sphynx?
- [Connections](https://docs.sphynx.sh/connections.md): The credentials your evals use
- [Variants](https://docs.sphynx.sh/evals/variants.md): The harness, model and sandbox a case runs on
- [Results](https://docs.sphynx.sh/evals/results.md): Read a batch, its runs and each trial
- [Run Claude Code evals](https://docs.sphynx.sh/guides/claude-code.md): Connect Claude once, check one trial, then compare models
- [Run evals in CI](https://docs.sphynx.sh/guides/ci.md): Gate pull requests on eval results
- [Eval an eve agent](https://docs.sphynx.sh/guides/eve.md): Run an agent built on eve as a command harness
- [Harness profiles](https://docs.sphynx.sh/evals/profiles.md): Version the files and settings you layer on a harness
- [Command harness](https://docs.sphynx.sh/evals/command-harness.md): Run your own agent as a process inside the sandbox
- [Run evals locally](https://docs.sphynx.sh/evals/local.md): Run an eval on your machine so it can reach local services
- [Import a case file](https://docs.sphynx.sh/evals/import.md): Turn existing JSON or YAML cases into a TypeScript suite
- [Sphynx CLI](https://docs.sphynx.sh/tools/cli.md): Run eval suites and connect credentials from a terminal
- [Sphynx MCP](https://docs.sphynx.sh/tools/mcp.md): Start and read evals from an MCP client
- [Quickstart](https://docs.sphynx.sh/getting-started.md): Define a suite and run it on a coding agent
- [Suites and cases](https://docs.sphynx.sh/evals/cases.md): Define what the agent does and how it passes
- [Mocks](https://docs.sphynx.sh/evals/mocks.md): Give agents typed local MCP servers, CLIs and HTTP APIs
- [Model judges](https://docs.sphynx.sh/evals/judges.md): Score answers with a model you choose
- [Conversations](https://docs.sphynx.sh/evals/conversations.md): Test an agent over a conversation, not a single prompt
- [Run evals from code](https://docs.sphynx.sh/guides/run-from-code.md): Start a batch from your own code and act on the result
- [API client](https://docs.sphynx.sh/sdk/client.md): Start batches and read runs from TypeScript
- [API overview](https://docs.sphynx.sh/api-reference/introduction.md): Call the Sphynx API over HTTP
- [Show which organization a credential uses](https://docs.sphynx.sh/api-reference/auth/show-which-organization-a-credential-uses.md): Runs, suites, and connectors started with this credential land in this organization.
- [List the integrations that can be connected](https://docs.sphynx.sh/api-reference/connectors/list-the-integrations-that-can-be-connected.md): Each integration states the auth methods it accepts and the fields each one needs.
- [List connected integrations](https://docs.sphynx.sh/api-reference/connectors/list-connected-integrations.md): The stored secret is never returned, only that a connection exists.
- [Connect an integration](https://docs.sphynx.sh/api-reference/connectors/connect-an-integration.md): Values are encrypted on arrival. The response states the connection without them.
- [Remove a connector](https://docs.sphynx.sh/api-reference/connectors/remove-a-connector.md)
- [Start a batch for a suite](https://docs.sphynx.sh/api-reference/batches/start-a-batch-for-a-suite.md): Starts one run per case and variant, and returns the batch while trials continue in the background. The command harness needs a profile with a run command; other harnesses take no run. Any harness may take an install command.
- [Get a batch](https://docs.sphynx.sh/api-reference/batches/get-a-batch.md)
- [List batches](https://docs.sphynx.sh/api-reference/batches/list-batches.md): Newest first. Pass the `next` cursor from a response to read the page after it; a null `next` means there are no more.
- [List suites](https://docs.sphynx.sh/api-reference/suites/list-suites.md): Most recently active first. Pass the `next` cursor from a response to read the page after it; a null `next` means there are no more.
- [Get a suite and the setup its cases share](https://docs.sphynx.sh/api-reference/suites/get-a-suite-and-the-setup-its-cases-share.md)
- [List cases](https://docs.sphynx.sh/api-reference/cases/list-cases.md)
- [Get a case, its variants and its versions](https://docs.sphynx.sh/api-reference/cases/get-a-case-its-variants-and-its-versions.md)
- [Start a batch for a case](https://docs.sphynx.sh/api-reference/cases/start-a-batch-for-a-case.md): Runs the case's newest version on the variants named, or on every variant it has run on, so they can be compared.
- [List a case's runs](https://docs.sphynx.sh/api-reference/runs/list-a-cases-runs.md): Newest first, 20 to a page.
- [Get a run and its trials](https://docs.sphynx.sh/api-reference/runs/get-a-run-and-its-trials.md)
- [List models available to a harness](https://docs.sphynx.sh/api-reference/models/list-models-available-to-a-harness.md)
- [Resolve a prompt](https://docs.sphynx.sh/api-reference/prompts/resolve-a-prompt.md): Returns the content a caller should send to a model. With no selector this follows the organization's default channel.
- [List prompts](https://docs.sphynx.sh/api-reference/prompts/list-prompts.md): Up to 100 prompts in the organization, without content.
- [Create a prompt](https://docs.sphynx.sh/api-reference/prompts/create-a-prompt.md): Creates the prompt and its first version in one call.
- [Add a version](https://docs.sphynx.sh/api-reference/prompts/add-a-version.md): Content is versioned, so updating a prompt appends a version rather than overwriting one. Earlier versions stay readable.
- [Promote a version to a channel](https://docs.sphynx.sh/api-reference/prompts/promote-a-version-to-a-channel.md): Points a channel, such as production, at a version. This is how a version goes live without callers changing anything.

## OpenAPI Specs

- [openapi](/openapi.json)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.