> ## Documentation Index > Fetch the complete documentation index at: https://docs.crewship.ai/llms.txt > Use this file to discover all available pages before exploring further. # crewship provider > Inspect the LLM providers this build can talk to, see how the agent sidecar routes each one, and send one live completion to prove a backend answers. # crewship provider Reads the LLM provider registry compiled into the binary, and — with `check` — dials a provider once to prove the wiring end to end. ```bash theme={null} crewship provider [flags] ``` Every subcommand is **local**. There is no server route behind them: the registry, the model catalog and the sidecar's routing table are all compiled in, so they work with no config file, no token, no workspace and no reachable server. `check` reaches the provider directly from the machine you run it on, which means it can test a local backend the server itself cannot see. Background on what these tables are: [Multi-provider LLM configuration](/guides/multi-provider-llm). ## Subcommands ### list ```bash theme={null} crewship provider list [--all] ``` **Flags:** | Flag | Type | Description | | ------- | ------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `--all` | `bool` | Also list providers that exist only in the embedded model catalog. Those have no codec in this build — they can be listed and priced, but nothing here can call them. | **Example:** ```bash theme={null} crewship provider list ``` ``` PROVIDER NAME CODEC AUTH KEY ENV KEY ENDPOINT MODELS anthropic Anthropic anthropic-messages x-api-key ANTHROPIC_API_KEY unset https://api.anthropic.com/v1/messages 13 openai OpenAI openai-compat bearer OPENAI_API_KEY unset https://api.openai.com/v1/chat/completions 47 ollama Ollama ollama-native none — not needed http://localhost:11434 0 ``` Rows are in registry **declaration order**, which is also the order the admin console's provider picker renders. Columns: | Column | Meaning | | ---------- | ----------------------------------------------------------------------------------------------------------------------------------- | | `PROVIDER` | The registry id. This is what `crewship keeper aux set --provider` accepts, and the `/` half of every paymaster rate key. | | `CODEC` | The wire format: `anthropic-messages`, `openai-compat`, or `ollama-native`. | | `AUTH` | How the credential reaches the wire: `x-api-key`, `bearer`, or `none`. | | `KEY ENV` | The environment variable the server reads a missing key from. | | `KEY` | Whether that variable is **set** — never its value. `not needed` is a provider that takes no credential. | | `ENDPOINT` | Where calls go. Hosted providers are fixed; `ollama` follows `KEEPER_OLLAMA_URL`. | | `MODELS` | How many entries the embedded catalog has for this provider. Ollama's `0` is correct — its model ids are whatever you pulled. | `--all` appends the catalog-only providers (`amazon-bedrock`, `deepseek`, `google`, `mistral`, `openrouter`, `xai`), sorted, with a dash in every column but `PROVIDER`, `NAME` and `MODELS`. That dash is the point: `deepseek` is listed because the catalog **prices** it, not because an evaluator slot can select it. `--format json` adds `registered`, `key_required`, `key_set` and `catalog_id` per row, which is the shape to script against: ```bash theme={null} crewship provider list --format json | jq '.providers[] | select(.key_set | not)' ``` ### route ```bash theme={null} crewship provider route list crewship provider route show ``` `list` answers the other half of the provider question. `provider list` says which providers **this binary** can construct and whether their key is in your environment; `route list` says how an **agent's** CLI reaches a provider through its sidecar — the path it dials on `127.0.0.1:9119`, where the sidecar forwards it, and which header or query parameter the credential is written into on the way out. The two tables are deliberately different sets. A provider here does not have to be one crewshipd can call, and vice versa. **Example:** ```bash theme={null} crewship provider route list ``` ``` ╭───────────────┬────────────────────┬───────────────────────────────────┬───────────────────────────────────────┬────────────┬──────────────────────────╮ │ PROVIDER │ PATH │ UPSTREAM │ AUTH │ CREDENTIAL │ BILLED AS │ ├───────────────┼────────────────────┼───────────────────────────────────┼───────────────────────────────────────┼────────────┼──────────────────────────┤ │ ANTHROPIC │ /v1 │ api.anthropic.com │ header x-api-key (+1 by token prefix) │ optional │ anthropic │ │ OPENAI │ /openai │ api.openai.com │ header Authorization: Bearer │ optional │ openai │ │ GOOGLE │ /gemini │ generativelanguage.googleapis.com │ header x-goog-api-key + query key │ optional │ google │ │ OPENROUTER │ /llm/openrouter │ openrouter.ai/api/v1 │ header Authorization: Bearer │ required │ openrouter │ │ OPENAI_COMPAT │ /llm/openai-compat │ from credential │ header Authorization: Bearer │ required │ openai-compat (unpriced) │ ╰───────────────┴────────────────────┴───────────────────────────────────┴───────────────────────────────────────┴────────────┴──────────────────────────╯ ``` | Column | Meaning | | ------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `PROVIDER` | The descriptor id. This is the value stored in a credential's `provider` column, and what `crewship credential create --provider` takes. | | `PATH` | The path prefix an agent's CLI is pointed at inside the container. | | `UPSTREAM` | The fixed dial target, base path included — or `from credential` when the endpoint is supplied by the credential itself. | | `AUTH` | Where the credential is written. A provider that accepts more than one token shape shows the default slot plus how many prefix-selected branches sit ahead of it; `route show` prints them all. | | `CREDENTIAL` | `required` refuses the call with a 503 when the sidecar holds no credential. `optional` forwards the agent's own request untouched, which is how an Anthropic OAuth token — which the sidecar never holds — still reaches the upstream. | | `BILLED AS` | The provider key cost-ledger rows are written under. `(unpriced)` means nothing prices that key, so calls through the route bill \$0. | `show` prints one route in full, including every auth branch: ```bash theme={null} crewship provider route show openrouter ``` ``` Provider: OPENROUTER Name: OpenRouter Agent path: /llm/openrouter Prefix: stripped before forwarding Upstream: openrouter.ai/api/v1 Credential: required Auth: header Authorization: Bearer Billed as: openrouter Usage parsed as: openai Key env vars: OPENROUTER_API_KEY ``` A provider with more than one token shape lists each branch — `provider route show anthropic` prints an `Auth (token sk-ant-oat…)` line above its `Auth (default)` one, which is how an OAuth token ends up in `Authorization` and an API key in `x-api-key` without either being a special case in the router. Like `list` and unlike `check`, both are **local and read-only**: the descriptor table is compiled in, nothing reads the credential vault, and no server is contacted. An unknown provider exits `3`: ```bash theme={null} crewship provider route show bedrock # unknown provider route "bedrock" (known: ANTHROPIC, OPENAI, GOOGLE, OPENROUTER, OPENAI_COMPAT) # exit 3 ``` Bedrock is not in that list and is not supported: it authenticates with SigV4 request signing, which the sidecar does not do. To create a credential for one of these providers, see [`crewship credential create`](/cli/credential#crewship-credential-create) — and for a `from credential` upstream, [an OpenAI-compatible endpoint](/cli/credential#an-openai-compatible-endpoint). ### check ```bash theme={null} crewship provider check --provider --model [flags] ``` Builds the provider for real, sends **one** completion, and prints the endpoint it dialled, the latency, the token counts the backend reported and what paymaster will bill for them. **Flags:** | Flag | Type | Description | | ------------ | ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `--provider` | `string` | A registry id (`anthropic`, `openai`, `ollama`) or an OpenAI-compatible preset key (`openai`, `deepseek`, `ollama-openai`, `vllm`). | | `--model` | `string` | Model id to send the completion to. | | `--base-url` | `string` | Override the endpoint. Required for a preset that ships none (`vllm`), and what makes a keyless local check possible. | | `--api-key` | `string` | API key. A **registry** provider falls back to its key environment variable (`ANTHROPIC_API_KEY`, `OPENAI_API_KEY`) and so does the `openai` preset; the other presets have none, so `deepseek` and a keyed `vllm` need this flag. Read it from your own environment (`--api-key "$DEEPSEEK_API_KEY"`) rather than pasting it: an argument is visible to every process on the host through `ps`. | | `--prompt` | `string` | Prompt to send. Defaults to `Reply with the single word: pong.` | | `--timeout` | `duration` | Deadline for the whole call. Default `1m0s`. | **Example — a local backend, no key:** ```bash theme={null} crewship provider check \ --provider ollama-openai \ --base-url http://localhost:11434/v1 \ --model qwen2.5:0.5b ``` ``` ollama-openai qwen2.5:0.5b codec openai-compat endpoint http://localhost:11434/v1 api key none pricing key ollama (what the ledger bills this as) latency 15.492s stop reason end_turn tokens in 37 out 4 cached-in 0 cache-write 0 cost $0.000000 (free rates: in $0.0000 out $0.0000 cached-in $0.0000 cache-write $0.0000) reply PONG! ``` `pricing key` is the line to read twice. It is `Provider.Name()`, not the preset key you passed — `ollama-openai` bills as `ollama` and `vllm` bills as `local`, which is how a self-hosted backend lands on a free rate row. **Zero tokens on a backend that clearly did work is the finding.** It means the response carried no `usage` block, and every call through that backend will be priced at \$0. On a streamed OpenAI-compatible endpoint the usual cause is a config without `IncludeUsage` — see [Multi-provider LLM configuration](/guides/multi-provider-llm#codecs). **Exit codes** follow the usual [CLI contract](/cli/overview). The two you will meet: ```bash theme={null} crewship provider check --provider openrouter --model qwen/qwen3-coder-flash # unknown provider "openrouter" (known: anthropic, openai, ollama, deepseek, ollama-openai, vllm) # exit 3 crewship provider check --provider deepseek --model deepseek-chat # deepseek check failed after 327ms: invalid DeepSeek API key # exit 7 ``` The vocabulary in the first message is generated from the registry and the preset table, so it cannot drift from what the command accepts. The second is the upstream's own refusal, preserved verbatim rather than flattened into "provider error". ## Related * [Multi-provider LLM configuration](/guides/multi-provider-llm) — what the registry and the codecs are. * [`crewship model`](/cli/model) — which model ids exist, and what they cost. * [Paymaster](/guides/paymaster) — the rate card `check` reports against. * [`crewship keeper`](/cli/keeper) — where a provider id is actually selected.