microsoft · Salil Das (sadlilas) · Aug 18, 2026 · Aug 18, 2026 · Aug 18, 2026
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,6 +7,18 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- **ChatGPT provider.** `provider.module: "openai-chatgpt"` is now a valid host-config value,
+  backed by `amplifier-module-provider-openai-chatgpt`. It uses a ChatGPT Plus/Pro/Team
+  subscription as the backend instead of a per-token API key, talking to the ChatGPT backend
+  (Codex API) rather than the public OpenAI API. Default model is `gpt-5.5`.
+  Auth is OAuth device-code, not an environment variable: the provider module drives an
+  interactive login at mount time (`login_on_mount`, default true) and caches tokens to
+  `~/.amplifier/openai-chatgpt-oauth.json`, refreshing them itself. Requires "Sign in with
+  device code" enabled in the account's ChatGPT Security settings. Like `github-copilot`,
+  `auth set openai-chatgpt` is refused — there is no static key to store.
+
 ## [0.12.0] — 2026-07-29
 
 ### Added

diff --git a/README.md b/README.md
@@ -43,7 +43,7 @@ Public integrations run opencode, paperclip, and NanoClaw on it: see [who has in
 
 `amplifier-agent` ships with:
 
-- Five providers behind one interface: Anthropic, OpenAI, Azure OpenAI, Ollama, and GitHub Copilot, with credentials read from the environment
+- Six providers behind one interface: Anthropic, OpenAI, Azure OpenAI, Ollama, GitHub Copilot, and ChatGPT (a Plus/Pro/Team subscription via OAuth device-code, no API key), with credentials read from the environment or a cached OAuth session
 - Role-based model routing, so a sub-agent gets a model matched to its job rather than the frontier model for everything, re-matched when you switch providers
 - Context management that keeps long sessions running, compacting history before it overruns the window
 - Tools for filesystem, bash, web, search, todo, and MCP

diff --git a/docs/CONFIGURATION.md b/docs/CONFIGURATION.md
@@ -11,6 +11,12 @@ Provider is auto-detected from environment variables in this precedence:
 3. `AZURE_OPENAI_API_KEY` + `AZURE_OPENAI_ENDPOINT`
 4. `OLLAMA_HOST` (defaults to `http://localhost:11434`)
 
+`github-copilot` and `openai-chatgpt` are excluded from this auto-detect chain -- neither resolves
+from a single API-key environment variable. `github-copilot` reads its own token chain from the
+environment (see below). `openai-chatgpt` has no credential env var at all: it authenticates via
+OAuth device-code, caching tokens to `~/.amplifier/openai-chatgpt-oauth.json`. Both must be
+selected explicitly with `provider.module` in a host config file.
+
 Override by passing `--config <path>` at a host config file that names a provider explicitly.
 
 > **Deprecated alias:** `AZURE_OPENAI_KEY` (without `_API_`) is still accepted as a fallback for backwards compatibility and triggers a one-time stderr warning when used. Prefer `AZURE_OPENAI_API_KEY`. The legacy name will be removed in a future release.
@@ -58,6 +64,8 @@ Resolution is **env-first**, so existing shell-rc workflows keep working unchang
 This matters for hosts that spawn `amplifier-agent` as a subprocess: once you have run `auth set` a single time, every subsequent invocation picks the key up automatically, from any terminal, from any directory, with or without exported environment variables.
 
 > **`github-copilot` is environment-only.** The other providers receive their credential through the mount config, so `auth set` works for them. The Copilot provider reads its token directly from the environment and ignores the config value, so `auth set github-copilot` is refused rather than storing a token the provider can never see. Set one of these instead (first non-empty wins): `COPILOT_AGENT_TOKEN`, `COPILOT_GITHUB_TOKEN`, `GH_TOKEN`, `GITHUB_TOKEN`.
+
+> **`openai-chatgpt` has no static key to store, so `auth set openai-chatgpt` is also refused.** It authenticates via OAuth device-code instead: the provider module drives an interactive login at mount time (`login_on_mount`, default true) and caches tokens to `~/.amplifier/openai-chatgpt-oauth.json`, refreshing them itself. Requires "Sign in with device code" enabled in the account's ChatGPT Security settings.
 >
 > ```bash
 > export GITHUB_TOKEN=$(gh auth token)

diff --git a/docs/INTEGRATION.md b/docs/INTEGRATION.md
@@ -8,7 +8,7 @@ The engine runs **one turn per invocation** and exits. Continuity across turns c
 
 `amplifier-agent` is a standalone binary. You do not need the Amplifier CLI, bundles, or any other repository in the `microsoft/amplifier*` family, and none of them is a substitute for it here.
 
-Use it when your software needs to run an agent: a loop with tools, file access, sub-agents, and/or multi-turn state. It also works for plain LLM calls, where you get routing across five providers behind one interface.
+Use it when your software needs to run an agent: a loop with tools, file access, sub-agents, and/or multi-turn state. It also works for plain LLM calls, where you get routing across six providers behind one interface.
 
 Then pick a surface below, install the engine ([INSTALL.md](INSTALL.md)), and finish with the [checklist](#checklist-for-a-new-integration).
 

diff --git a/docs/LAYERS_AND_RELEASES.md b/docs/LAYERS_AND_RELEASES.md
@@ -115,7 +115,7 @@ The engine ships with `bundle.md` baked into the wheel. It declares which module
 
 **Pre-wired modules:**
 
-- **Providers:** `provider-anthropic`, `provider-openai`, `provider-azure-openai`, `provider-ollama`, `provider-github-copilot`
+- **Providers:** `provider-anthropic`, `provider-openai`, `provider-azure-openai`, `provider-ollama`, `provider-github-copilot`, `provider-openai-chatgpt`
 - **Orchestrator:** `loop-streaming` (with `extended_thinking: true`)
 - **Context:** `context-simple` (300K tokens, auto-compact at 80%)
 - **Tools:** `tool-filesystem`, `tool-bash`, `tool-web`, `tool-search`, `tool-todo`, `tool-apply-patch`, `tool-delegate`, `tool-mcp`, `tool-skills`, `tool-mode`, `tool-recipes`

diff --git a/docs/architecture/architecture.dot b/docs/architecture/architecture.dot
@@ -84,7 +84,7 @@ digraph amplifier_agent {
 
     node [fillcolor="#d9d9d9"];
     kernel    [label="amplifier-foundation\nbundle / session kernel"];
-    providers [label="LLM providers\nanthropic · openai · azure\nollama · copilot"];
+    providers [label="LLM providers\nanthropic · openai · azure\nollama · copilot · chatgpt"];
     mcp       [label="MCP servers"];
   }
 

diff --git a/docs/spec/bundle-and-cache.md b/docs/spec/bundle-and-cache.md
@@ -30,7 +30,7 @@ default_provider: anthropic                REQUIRED, engine-level, top-level key
 
 providers:                                 install-only stubs, no config and no credentials
   provider-anthropic, provider-openai, provider-azure-openai,
-  provider-ollama, provider-github-copilot
+  provider-ollama, provider-github-copilot, provider-openai-chatgpt
 
 session.orchestrator: loop-streaming       extended_thinking: true
 session.context:      context-simple       max_tokens 300000, auto_compact

diff --git a/docs/spec/cli.md b/docs/spec/cli.md
@@ -197,7 +197,8 @@ serve restart
 auth set PROVIDER [API_KEY] [--stdin] [--endpoint URL]
     Writes ~/.amplifier-agent/credentials.json (mode 0600, atomic write). --stdin reads the key
     from stdin so it never appears in argv. --endpoint carries an Azure-style deployment URL.
-    `github-copilot` is refused: it reads its token from the environment.
+    `github-copilot` is refused: it reads its token from the environment. `openai-chatgpt` is
+    also refused: it has no static key, authenticating instead via OAuth device-code.
 
 auth list
     Per-provider table: masked value plus source (`env=<VAR>` / `file` / `default` / `not set`).

diff --git a/docs/spec/host-config.md b/docs/spec/host-config.md
@@ -81,7 +81,7 @@ approval.patterns  must be a list of strings
 `provider` selects the provider module and carries its config.
 
 ```
-provider.module   one of: anthropic, openai, azure-openai, ollama, github-copilot
+provider.module   one of: anthropic, openai, azure-openai, ollama, github-copilot, openai-chatgpt
 provider.config   free-form; belongs to the provider module
 ```
 
@@ -101,7 +101,7 @@ module config. Closed per-entry schema:
 }
 ```
 
-`module` defaults to the entry's own id when omitted and must be one of the five valid module names.
+`module` defaults to the entry's own id when omitted and must be one of the six valid module names.
 `config` must be an object. Unknown keys inside an entry raise `config_unknown_key`. An empty
 `providers` object passes validation; HTTP startup rejects it separately at boot so single-turn mode
 never trips on a stale block.

diff --git a/docs/spec/providers-and-models.md b/docs/spec/providers-and-models.md
@@ -9,7 +9,7 @@ routes a wire `model` field to a provider (see `http-face.md`).
 
 ## Supported providers
 
-Five providers are supported, and only five. The provider name is the value used in configuration,
+Six providers are supported, and only six. The provider name is the value used in configuration,
 in `auth` subcommands, and in `models list --provider`.
 
 ```
@@ -18,10 +18,11 @@ openai          provider-openai
 azure-openai    provider-azure-openai
 ollama          provider-ollama
 github-copilot  provider-github-copilot
+openai-chatgpt  provider-openai-chatgpt
 ```
 
 Each module is installed from `git+https://github.com/microsoft/amplifier-module-<module>@main`.
-All five are declared by the shipped bundle as install-only, so preparing the bundle makes every
+All six are declared by the shipped bundle as install-only, so preparing the bundle makes every
 provider importable before any session exists.
 
 The agent holds no static table of default models, credential field shapes, or display names. Those
@@ -48,6 +49,7 @@ openai          OPENAI_API_KEY
 azure-openai    AZURE_OPENAI_API_KEY, then AZURE_OPENAI_KEY
 ollama          OLLAMA_HOST, then OLLAMA_BASE_URL
 github-copilot  GITHUB_TOKEN
+openai-chatgpt  (none -- OAuth device-code)
 ```
 
 `AZURE_OPENAI_KEY` is the only deprecated alias. Consulting it emits a one-time warning on stderr.
@@ -61,6 +63,11 @@ github-copilot lists only `GITHUB_TOKEN` here. The provider module resolves its
 (`COPILOT_AGENT_TOKEN`, `COPILOT_GITHUB_TOKEN`, `GH_TOKEN`, `GITHUB_TOKEN`); listing those here
 would mark them deprecated, which they are not.
 
+openai-chatgpt has no environment variable at all. It resolves from a cached OAuth token file
+(`~/.amplifier/openai-chatgpt-oauth.json`), written by the provider module's own device-code login
+flow (`login_on_mount`) and refreshed automatically thereafter. Its resolution reports source
+`"file"` when a token is cached and `"none"` otherwise -- never `"env"`.
+
 A resolution reports the provider, whether it resolved, the source (`env`, `file`, `default`, or
 `none`), the variable consulted, and the resolved fields. Ollama backed only by the built-in default
 host reports unresolved on purpose, so auto-enrollment does not enlist a local daemon that may not
@@ -100,9 +107,13 @@ on the next write. Unknown provider keys round-trip verbatim. A malformed file f
 with an error but resolves as empty on the read path, so one bad write does not break every later
 invocation.
 
-`auth set github-copilot` is refused. The agent normalizes every credential into an `api_key`
-config field, and that provider's module reads only the environment, so a stored value would report
-success and change nothing. This refusal is temporary and specific to that one provider.
+`auth set github-copilot` and `auth set openai-chatgpt` are both refused, for different reasons.
+The agent normalizes every credential into an `api_key` config field: github-copilot's module reads
+only the environment and ignores it, so a stored value would report success and change nothing.
+openai-chatgpt has no static key at all -- it authenticates via OAuth device-code and caches tokens
+to `~/.amplifier/openai-chatgpt-oauth.json`, refreshed by the provider module itself. Both refusals
+are enumerated in the same `_CONFIG_CREDENTIAL_UNSUPPORTED` gate; this is temporary and specific to
+these two providers.
 
 `auth clear` without `--force` exits 2.
 
@@ -114,7 +125,7 @@ success and change nothing. This refusal is temporary and specific to that one p
 3. no further fallback: a bundle declaring neither is a hard error at boot
 ```
 
-`provider.module` is closed to the five supported names. Any other value fails validation with
+`provider.module` is closed to the six supported names. Any other value fails validation with
 error code `config_invalid_provider_module`. `"auto"` is not a valid value.
 
 There is no `--provider` flag and no environment-based provider auto-detection. See Non-goals.

diff --git a/skills/amplifier-agent/SKILL.md b/skills/amplifier-agent/SKILL.md
@@ -22,7 +22,7 @@ metadata:
 
 `amplifier-agent` is an agent engine that other software runs on. Give it a prompt and it runs the full loop, with tools, sub-agents, skills, and MCP, then returns a result. Anything that can spawn a subprocess can use it; Python hosts can embed the engine library in-process instead.
 
-Reach for it when the project needs an *agent* (a tool loop, file access, sub-agents, multi-turn state) rather than a single completion. You can also use it for plain LLM calls, with routing across five providers behind one interface.
+Reach for it when the project needs an *agent* (a tool loop, file access, sub-agents, multi-turn state) rather than a single completion. You can also use it for plain LLM calls, with routing across six providers behind one interface.
 
 **The engine runs one turn per invocation and exits.** Continuity across turns comes from a session id, not from a long-lived process. Every surface below is a different way of delivering a prompt to that same engine.
 
@@ -54,7 +54,7 @@ The installer needs `uv` and `curl` and will not bootstrap them silently; it tel
 
 Install as **the same user that runs the host process**; a host spawning a subprocess inherits that user's `PATH`. `amplifier-agent doctor` is the check that the install actually works, so run it before writing any integration code.
 
-Credentials are read from the environment, first match wins: `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `AZURE_OPENAI_API_KEY` plus `AZURE_OPENAI_ENDPOINT`, `OLLAMA_HOST`. GitHub Copilot is environment-only (`COPILOT_AGENT_TOKEN`, `COPILOT_GITHUB_TOKEN`, `GH_TOKEN`, `GITHUB_TOKEN`). Or store one with `amplifier-agent auth set anthropic sk-ant-...`.
+Credentials are read from the environment, first match wins: `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `AZURE_OPENAI_API_KEY` plus `AZURE_OPENAI_ENDPOINT`, `OLLAMA_HOST`. GitHub Copilot is environment-only (`COPILOT_AGENT_TOKEN`, `COPILOT_GITHUB_TOKEN`, `GH_TOKEN`, `GITHUB_TOKEN`). ChatGPT (`openai-chatgpt`) has no credential env var at all: it authenticates via OAuth device-code, caching tokens to `~/.amplifier/openai-chatgpt-oauth.json`. Or store a static key with `amplifier-agent auth set anthropic sk-ant-...` (not supported for github-copilot or openai-chatgpt).
 
 ## Pick a surface
 
@@ -191,7 +191,7 @@ A per-instance config file looks like this:
 | `config_unreadable`, `config_malformed_json` | The `--config` file could not be opened, or is not a JSON object | Check the path the host wrote, and that it serialized an object |
 | `config_unknown_key` | Unrecognized **top-level** config key | The top level is closed: `approval`, `provider`, `providers`, `mcp`, `skills`, `debug`, `allowProtocolSkew` |
 | `config_invalid_type` | A known key has the wrong type, or an unknown sub-key in a closed inner shape | `skills.*` and `debug.*` are closed and raise this rather than `config_unknown_key`, which is reserved for the top level and `providers.<id>` entries |
-| `config_invalid_provider_module` | `provider.module` is not a known provider | One of `anthropic`, `openai`, `azure-openai`, `ollama`, `github-copilot`. `"auto"` is not valid |
+| `config_invalid_provider_module` | `provider.module` is not a known provider | One of `anthropic`, `openai`, `azure-openai`, `ollama`, `github-copilot`, `openai-chatgpt`. `"auto"` is not valid |
 | `protocol_version_mismatch` | Wrapper and engine protocol versions differ | Update the lagging side. `allowProtocolSkew` is an unblock, not a fix |
 | `lifecycle_unsupported` | `submit()` called twice on one handle | New handle per turn, same `sessionId` with `resume` |
 | `env_injection_rejected` | The wrapper refused the environment you asked it to inject | Check the key against the wrapper's allowlist and blocked-key list |

diff --git a/src/amplifier_agent_cli/admin/auth.py b/src/amplifier_agent_cli/admin/auth.py
@@ -71,7 +71,7 @@
 #: agent-delivered credential from config first, then its env chain, then cached
 #: OAuth. When that lands, DELETE this constant and its gate in ``auth_set``
 #: outright. Do NOT grow it into a general provider-capability mechanism.
-_CONFIG_CREDENTIAL_UNSUPPORTED: Final[frozenset[str]] = frozenset({"github-copilot"})
+_CONFIG_CREDENTIAL_UNSUPPORTED: Final[frozenset[str]] = frozenset({"github-copilot", "openai-chatgpt"})
 
 
 # ---------------------------------------------------------------------------

diff --git a/src/amplifier_agent_cli/provider_sources.py b/src/amplifier_agent_cli/provider_sources.py
@@ -105,7 +105,14 @@ def _emit_legacy_env_var_notice(legacy_var: str, preferred_var: str) -> None:
 #: ``models list --provider <name>``, or aggregate iteration in admin
 #: commands) against the supported set. Kept in sync with
 #: ``PROVIDER_CATALOG.keys()``.
-KNOWN_PROVIDERS: Final[tuple[str, ...]] = ("anthropic", "openai", "azure-openai", "ollama", "github-copilot")
+KNOWN_PROVIDERS: Final[tuple[str, ...]] = (
+    "anthropic",
+    "openai",
+    "azure-openai",
+    "ollama",
+    "github-copilot",
+    "openai-chatgpt",
+)
 
 
 #: Map provider short-name → bootstrap catalog row.
@@ -136,6 +143,10 @@ def _emit_legacy_env_var_notice(legacy_var: str, preferred_var: str) -> None:
         "module": "provider-github-copilot",
         "source": "git+https://github.com/microsoft/amplifier-module-provider-github-copilot@main",
     },
+    "openai-chatgpt": {
+        "module": "provider-openai-chatgpt",
+        "source": "git+https://github.com/microsoft/amplifier-module-provider-openai-chatgpt@main",
+    },
 }
 
 
@@ -375,6 +386,42 @@ def resolve_credential_detailed(provider_name: str) -> CredentialResolution:
             fields={"host": _OLLAMA_DEFAULT_HOST},
         )
 
+    if provider_name == "openai-chatgpt":
+        # ChatGPT subscription provider (provider-openai-chatgpt): authenticates
+        # via OAuth device-code, NOT an api key. There is no credential env var;
+        # the module caches its OAuth tokens to a file and refreshes them itself.
+        # We report "resolvable" iff that token cache exists and parses with a
+        # token present -- an honest signal of whether a prior device-code login
+        # happened -- without ever reading or emitting the token material. The
+        # module's own ``login_on_mount`` drives the interactive device-code flow
+        # at mount time when no cache exists. ``auth set`` is refused for this
+        # provider (see _CONFIG_CREDENTIAL_UNSUPPORTED in admin.auth) since there
+        # is no static key to store.
+        import json
+        from pathlib import Path
+
+        token_file = Path("~/.amplifier/openai-chatgpt-oauth.json").expanduser()
+        try:
+            data = json.loads(token_file.read_text())
+            has_token = isinstance(data, dict) and bool(data.get("access_token") or data.get("refresh_token"))
+        except (OSError, ValueError):
+            has_token = False
+        if has_token:
+            return CredentialResolution(
+                provider=provider_name,
+                resolved=True,
+                source="file",
+                env_var=None,
+                fields={},
+            )
+        return CredentialResolution(
+            provider=provider_name,
+            resolved=False,
+            source="none",
+            env_var=None,
+            fields={},
+        )
+
     env_vars = PROVIDER_CREDENTIAL_VARS.get(provider_name)
     if not env_vars:
         return CredentialResolution(provider=provider_name, resolved=False, source="none", env_var=None, fields={})

diff --git a/src/amplifier_agent_lib/bundle/bundle.md b/src/amplifier_agent_lib/bundle/bundle.md
@@ -68,6 +68,8 @@ providers:
     source: git+https://github.com/microsoft/amplifier-module-provider-ollama@main
   - module: provider-github-copilot
     source: git+https://github.com/microsoft/amplifier-module-provider-github-copilot@main
+  - module: provider-openai-chatgpt
+    source: git+https://github.com/microsoft/amplifier-module-provider-openai-chatgpt@main
 
 session:
   raw: true