Atlas fails with ProviderNoProvidersError because zero providers autoloaded, and a provider only autoloads when its credential is present, either as an api key in Atlas auth storage or as the provider's env vars. Fix it by running `atlas auth login` and adding at least one provider credential, or by exporting the provider's API key env var in the shell that launches Atlas. Then run `atlas providers` and confirm at least one entry appears, and run `atlas models` to confirm that provider exposes models.
Why does Atlas report No providers configured
Atlas reports No providers configured because Provider.defaultModel raises NoProvidersError when 0 providers autoload. With 0 providers loaded, Atlas has no provider to choose a default model from, so the session cannot start a request, and nothing appears under `atlas providers`.
Autoloading in Atlas is credential-driven. A provider is loaded only when its credential is present, either as an api key stored in Atlas auth storage or as the provider's environment variables. No credential means no autoload, and zero autoloaded providers means Provider.defaultModel has nothing to pick from, so it raises NoProvidersError. The empty output of `atlas providers` is not a display bug. It is the accurate report of a session with no provider loaded. That distinction matters, because it tells you the fix is to supply a credential, not to change a model or edit a prompt.
How do I fix ProviderNoProvidersError in Atlas
Step 1 of the documented fix is to run `atlas auth login` and add at least 1 provider credential. Atlas autoloads a provider only when its credential is present, so a single valid credential is the minimum required to move Atlas out of the No providers configured state.
`atlas auth login` writes the credential into Atlas auth storage, which is exactly what the autoload check looks for. Add one provider first rather than trying to configure several at once, because ProviderNoProvidersError is a threshold error: it clears the moment one provider loads. Once `atlas auth login` completes, restart the Atlas session so autoload runs again. If the login flow succeeded but ProviderNoProvidersError persists, the credential is not reaching autoload, and the environment variable path described below is the alternative worth trying.
Can I configure an Atlas provider with an env var instead of atlas auth login
Yes. Step 2 of the documented fix is exactly that: export the provider's API key env var in the shell that launches Atlas. Atlas autoloads a provider when its credential is present as an api key in auth storage or as the provider's env vars, so the provider loads without ever running `atlas auth login`.
The environment variable path is the right choice for CI, containers, and any machine where you would rather not persist a credential to disk. The one requirement that trips people up is scope: the API key env var must be exported in the shell that launches Atlas. A variable set in a dotfile that the current shell never sourced, or set in a different terminal tab, is invisible to the Atlas process, so autoload still sees zero credentials and Provider.defaultModel still raises NoProvidersError. Export it, then launch Atlas from that same shell.
How to verify the No providers configured fix worked
Run `atlas providers` and confirm at least 1 entry appears. An empty list means 0 providers autoloaded and ProviderNoProvidersError will fire again. One or more entries means the credential reached autoload, and Provider.defaultModel now has a provider to choose a default model from.
`atlas providers` is the direct test of the condition that raises the error, which makes it the correct verification step. Confirm at least one entry appears there before you try anything else. After that, run `atlas models` to confirm that provider exposes models. The two commands answer two different questions in sequence: `atlas providers` answers whether the provider loaded, and `atlas models` answers whether the loaded provider actually has models available to it. Passing the first and failing the second is a different problem than No providers configured, and should be debugged separately.
Why atlas providers is empty after adding a credential
An empty `atlas providers` list after adding a credential means the credential is not where Atlas autoload looks. Atlas autoloads from exactly 2 places, an api key in auth storage or the provider's env vars, so a key written anywhere else leaves 0 providers loaded and NoProvidersError intact.
Two checks resolve almost every case. First, re-run `atlas auth login` and watch it complete, rather than assuming a previous run stored the key. Second, if you took the environment variable route, confirm the API key env var is exported in the very shell that launches Atlas. Being honest about the caveat: Atlas does not tell you which of the two sources it failed to find a credential in. The error text stops at No providers configured, so you have to check both paths yourself rather than reading a precise answer out of the message.
What No providers configured does not mean in Atlas
No providers configured in Atlas does not mean a model is unavailable, a request was rejected, or a quota was spent. Those are 3 separate failures. ProviderNoProvidersError is raised by Provider.defaultModel before any request leaves your machine, because 0 providers autoloaded in the first place.
The failure is upstream of everything a user usually reaches for. There is no model to change, because a model cannot be selected without a provider. There is no prompt to shorten, because no prompt was ever sent. There is no billing question, because no provider was contacted. Treat ProviderNoProvidersError as a setup error rather than a runtime error. Supply one credential through `atlas auth login` or through the provider's API key env var, confirm with `atlas providers`, and the rest of the session behaves normally from there.
How to fix it
- 01Run `atlas auth login` and add at least one provider credential.
- 02Or export the provider's API key env var in the shell that launches Atlas.
- 03Run `atlas providers` and confirm at least one entry appears.
- 04Then run `atlas models` to confirm that provider exposes models.
Frequently asked questions
- how to fix No providers configured in atlas
- Run `atlas auth login` and add at least one provider credential, or export the provider's API key env var in the shell that launches Atlas. Then run `atlas providers` and confirm at least one entry appears.
- what is ProviderNoProvidersError in atlas
- ProviderNoProvidersError is raised by Provider.defaultModel when zero providers autoload in Atlas. A provider autoloads only when its credential is present, as an api key in auth storage or as the provider's env vars.
- atlas providers returns nothing, what do I do
- An empty `atlas providers` list means no provider autoloaded. Add a credential with `atlas auth login`, or export the provider's API key env var in the shell that launches Atlas, then run `atlas providers` again.
- does atlas need an api key to start
- Atlas needs at least one provider credential before it can pick a default model. Without a credential in auth storage or in the provider's env vars, zero providers autoload and Atlas raises NoProvidersError.
- why doesn't atlas see my exported api key
- The provider's API key env var must be exported in the same shell that launches Atlas. A key set only in a dotfile the shell never sourced, or set in another terminal tab, is invisible to Atlas, so zero providers autoload.
- how do I confirm my atlas provider is working after login
- Run `atlas providers` and confirm at least one entry appears, then run `atlas models` to confirm that provider exposes models. The first command checks that the provider loaded, the second checks that it has usable models.
- will switching models fix No providers configured in atlas
- No. ProviderNoProvidersError fires because zero providers autoloaded, so there is no provider to select a model from. Add a provider credential with `atlas auth login` instead of changing models.
Try Atlas in your terminal
The terminal-native AI coding agent. Free core, single binary.
Install AtlasRelated guides
Atlas with DeepSeek-R1 (local via Ollama): Visible Reasoning at Every Size in 2026
DeepSeek-R1 runs locally from 1.5B to 671B under one Ollama tag, with a 128K to 164K context, free self-hosted, or $1.35 / $5.40 per Mtok via Bedrock. Atlas setup.
Atlas with AllenAI Olmo 3 32B Think: the fully open reasoning model in 2026
Run Atlas on AllenAI Olmo 3 32B Think in 2026: 65,536 token context, OpenRouter $0.15/$0.50 per Mtok, and the only weights, data, and training code you can audit.
Atlas with Kimi K2 Thinking: Open-Weights Reasoning at $0.60 per Mtok (2026)
Kimi K2 Thinking gives Atlas reasoning at $0.60 / $2.50 per Mtok on a 262,144 token window. Open weights, and it sustains the long tool-calling chains agents need.
Atlas vs Roo Code: Terminal AI Coding Agents in 2026
Comparing Atlas and Roo Code in 2026. Atlas offers terminal-native TUI, permission-gated tools, and diff review. Roo Code, a VS Code extension, shut down May 15, 2026.
Atlas with Llama 4 Scout: the 3.5M Token Context Model in 2026
Llama 4 Scout gives Atlas a 3.5M token context on Bedrock at $0.17 / $0.66 per Mtok, or $0.10 / $0.30 on DeepInfra. Setup, the portability trap, and when to switch.
Atlas for Godot: Terminal-Native AI Coding for GDScript and the Node Tree in 2026
Atlas is a terminal-native AI coding agent for Godot in 2026, working across GDScript, the node tree, and signals, where scene structure is half the program.
Atlas with DeepSeek Chat: 384,000 Token Output at $0.28 per Mtok in 2026
Run Atlas on DeepSeek Chat in 2026. DeepSeek's non-reasoning endpoint gives 1M tokens (1,000,000) of context at $0.14 per Mtok input, $0.28 per Mtok output.
Atlas for WebAssembly: Terminal-Native AI Coding in 2026
Atlas is a terminal-native AI coding agent for WebAssembly. Audit wasm-bindgen exports, batch calls across the JS boundary, and run wasm-pack test --node in 2026.