* docs(ai,coding-agent): generate models.json reference from schemas
Add npm run generate:references, which renders docs/models-configuration.md
from the models.json TypeBox schemas, and describe provider and model fields
in those schemas. Overrides get their own descriptions without defaults, and
compat sections cover the Responses API aliases.
closes#10790
* Modify changelog
* docs(ai,coding-agent): use shared thinkingLevelMap and promptCache schemas
Replace the per-use schema helpers with fixed schemas whose descriptions
hold for both models and modelOverrides entries.
* feat(coding-agent): default NVIDIA to Nemotron 3.5 Super VL
Co-authored-by: prademacher-nvidia <235594174+prademacher-nvidia@users.noreply.github.com>
* docs(coding-agent): changelog for NVIDIA default model change
* fix(ai): run NVIDIA E2E tests against Nemotron 3.5 Super VL
---------
Co-authored-by: prademacher-nvidia <235594174+prademacher-nvidia@users.noreply.github.com>
Compaction summaries were limited to 0.8 * compaction.reserveTokens (13107 by default), turn-prefix summaries to 0.5 * reserveTokens, and branch summaries to a hardcoded 4096. On providers where reasoning shares the output limit (OpenAI Responses, adaptive Anthropic), high thinking levels used most of that limit and compaction failed with "generation hit the token cap", even though the model allows far more output.
Tying the limit to reserveTokens assumed the summary request is as large as the full context, so its output had to fit in the reserved space. It is usually much smaller: recent messages kept by keepRecentTokens are not sent, a split turn goes to a separate request, tool results are truncated, and tool definitions and the normal system prompt are not included. Manual compaction can also run long before the context is full. The provider layer already sizes output from the actual request, using the model's output limit reduced to fit the remaining context, as for normal turns.
Summary requests now set no maxTokens and rely on that. compaction.reserveTokens only controls when compaction starts.
generateSummary() and generateSummaryWithUsage() now ignore the reserveTokens parameter.
closes#9512, closes#9075, closes#10908
* fix(coding-agent): resolve model references through one provider selection policy
Model references from --model, --models/enabledModels, /model, and auth credential printing now share one resolver. It collects every candidate before choosing a provider, so catalog order never decides, a :<thinking> suffix cannot change the provider, and an ambiguous exact match no longer falls through to a fuzzy match.
provider/id still prefers the named provider's exact or fuzzy match. A gateway exposing provider/id as a raw model id is used when the named provider has no match, or lacks credentials while the gateway has them. Several gateways without a unique authenticated one are reported as ambiguous.
fixes#10160
* docs(coding-agent): shorten model selection notes in CLI reference
* docs(coding-agent): condense #10160 changelog entry
* fix(coding-agent): move model reference resolution into model-resolver
* fix(coding-agent): warn when a model scope pattern matches several providers
* docs(coding-agent): trim model resolver comments to the reasons
* fix(coding-agent): build custom model ids from the resolver's not-found target
The SDK always passed initial tools, so the constructor never restored the
loadout recorded in the transcript, and resumed sessions dropped tools activated
during the session. Sessions that use the default tools now restore the recorded
loadout, with --tools +name/-name entries applied to it.
fixes#10785
A just-removed auth.json.lock directory can stay delete-pending on Windows
while another process holds a handle, making mkdir fail with EPERM. Retry
this briefly (capped separately from ELOCKED) so real permission errors
still fail fast.
fixes#10801
The codemode row shows "Elapsed" while the script runs and "Took" once
it finishes, like bash rows. The timer logic moves from the bash renderer
to render-utils, and scripts send an initial update so the timer starts
before the first nested call.
closes#10708
Entries like !~/skills/pdf, +~/x and -~/x in the extensions, skills,
prompts, and themes settings were compared with a literal ~ and never
matched.
fixes#10765
On each OpenRouter refresh, combine the catalog (built-in models plus a newer pi.dev catalog) with GET /models/user for the user's key: keep only the listed chat models and take their context size, max output and prices from it. Image and classifier models are kept. The combined list is saved and replaces the OpenRouter catalog, including offline.
- ai: add personalizeOpenRouterCatalog and RefreshModelsContext.baseUrl, so the request follows a base URL configured in models.json (regional endpoints)
- ai: share OpenRouter price conversion with the generator; negative prices (router models such as openrouter/auto) become 0
- coding-agent: add withPersonalizedCatalog next to withRemoteCatalog
- docs: OpenRouter section in providers.md
fixes#10353
Runs started by pi.sendMessage(..., { triggerTurn: true }) skipped before_agent_start. After their first tool call, the next-turn refresh fell back to the base prompt options and patched away the sections handlers had added, changing the system prompt mid-run and breaking the provider prompt cache.
These runs now go through the same setup as prompt(): compaction check, before_agent_start, tool loadout, nextTurn messages, handler messages, and the system patch. BeforeAgentStartEvent is a union on messageType ("user" or "custom"); custom events carry the starting message.
The session counts as busy from the start of setup, so messages sent while handlers run go into the starting run instead of starting a second one, and an abort requested meanwhile aborts the run once it starts.
closes#10267, closes#5581, closes#10639, closes#10289
partly fixes#10606
Tools that reach other tools declare it with reachesTools(input); MCP uses it to wait for the servers a call needs and to decide which codemode or tool_search tool to activate, instead of parsing codemode scripts and comparing parameter schemas. Namespaces get aliases, replacing codemode's MCP-specific __ suffix rule.
pi-codemode no longer knows MCP: MCP_TYPESCRIPT_PREAMBLE and mcpStructuredContentSchema() are removed, and the MCP result schema describes itself, with the server's local $refs relocated so they resolve as before.
Replayed thinking is often not part of the live context (for example,
radius drops it, and Anthropic drops blocks that do not match their
request), but serializeConversation included all of it. A Pi Mobile
session at ~996K of 1M context tokens produced 1.38M-token summarization
requests that failed on every attempt. Providers only return summaries of
the reasoning anyway; visible text, tool calls, and tool results record
what happened.
The nested worktree dedup compared the context file's realpath with the
main repo's file, so a worktree AGENTS.md symlinked to the main repo's
copy was skipped along with it. Compare the walked directory's realpath
plus the file name instead.
fixes#10681
Report idle, working, blocked, done, and error to terminals that support
the Program Status Protocol. ProcessTerminal sends the support query with
the keyboard query and reports only after the terminal answers before DA1;
PI_PROGRAM_STATUS=1|0 overrides detection. Interactive mode reports runs,
compaction, extension dialogs, and logins. agent_settled now carries
aborted so cancelled runs report idle.
closes#10607
Adds the openai-decisions classifier API with gpt-6-luna on the openai provider. bool questions map to predicates with the criteria in the instructions, refusals fail the result, and gateway timeouts (504) are not retried. The model needs an API key and is hidden from availability while openai uses Sign in with ChatGPT.
Adds optional images to ClassifierContext. Models without image input and APIs that cannot send images return an error result. Codemode's models.classify() accepts images.
Moves the shared classifier HTTP code out of system-one-shared.ts into classifier-shared.ts.
pi-ai's AssistantMessageEventStream sets AssistantMessage.durationMs with a
monotonic clock for every API implementation. The agent loop and the durable
tool task set ToolResultMessage.durationMs around execute(). Durable task
records carry startedAt/endedAt stamped by the Session. The coding agent's bash
renderer shows Took from the recorded duration, also after a reload.
closes#10549
A --tools list of only +name/-name entries changes the default tool
selection like the defaultTools setting, e.g. pi -t +codemode. Mixed
lists and patterns in modifiers are rejected, and /reload keeps tools
removed with -name removed.
Rules and the tool list follow the declared tools (active minus
hiddenDeclarations). The skills hint names no tool when the reader is
hidden. Codemode shows each tool's prompt guidelines with its
declaration, via the new ToolLoadout.getPromptGuidelines().
closes#10343
--tools no longer removes MCP tools unless an entry starts with mcp__, so
pi --tools codemode can reach MCP servers. Unnamed MCP tools are never
declared directly; only tool_search can load them. --no-tools still
removes everything.
--tools and --exclude-tools accept * patterns, for example
--tools read,codemode,'mcp__radius__*'. --no-mcp disables the built-in
MCP support for one run.
* feat(ai): support Azure Foundry Chat Completions deployments
The Azure provider only implemented the Responses API, so Foundry
deployments that speak Chat Completions could not be used. On Chat
Completions the deepseek/deepseek-v4-pro model sends `thinking` and
`prompt_cache_key`, both of which Azure rejects with 400.
The provider now dispatches openai-completions alongside
azure-openai-responses. A stream wrapper resolves the Azure endpoint
onto the model first, since the shared openai-completions
implementation reads it off the model.
The request-shape differences are catalog compat rather than api code:
reasoning_effort instead of DeepSeek's thinking field, no prompt cache
parameters, reasoning_content kept on assistant turns so the prefix
Foundry cached stays byte-identical, and effort clamped to the
low/medium/high the deployment accepts.
The provider is renamed from azure-openai-responses to azure, since it
now serves more than the Responses API. The azure-openai-responses api
id is unchanged. Breaking: auth.json and models.json entries keyed on
azure-openai-responses must move to azure, and existing sessions see a
different provider on their next turn.
closes#9645
* feat(ai): apply the Azure deployment name map to Chat Completions
The Responses path already sent the deployment name from
AZURE_OPENAI_DEPLOYMENT_NAME_MAP. The Chat Completions path sent the
catalog model id, so any deployment not named after the model 404'd.
The Azure provider now swaps the deployment name into the request via
onPayload before dispatch, then runs the caller's own onPayload on it.
model.id stays the catalog id. The map helpers move to
azure-openai-config.ts so the provider can share them without
importing the SDK.
* docs(ai,coding-agent): add changelog entries for the Azure provider rename
Breaking: the azure-openai-responses provider is now azure. Users rename
the key in auth.json, models.json, and settings.json; SDK users switch to
getModel("azure", ...) and providers/azure.
Also note Azure Foundry Chat Completions support (DeepSeek V4 Pro).
---------
Co-authored-by: David Brailovsky <davidbrai@gmail.com>
image() output is also written to a temp file and the result names the
path before each image, so later turns can reference generated images.
Output files are now created through one helper (still tmpdir()) and
are readable only by the user.
closes#10310
Add pi.registerToolRenderer((toolName, next) => renderers), which chooses the
renderers for calls to any tool, including tools that are not registered.
Resolvers run in extension load order; next() returns what the remaining
resolvers, then the registered tool, would use. Interactive mode and HTML
export both resolve through it.
The MCP extension uses it to render every mcp__<server>__<tool> call with the
MCP renderers, so calls in a resumed session no longer render fully expanded
before their server connects, or for good if it never does.
closes#10285
Adds oauth.clientRegistration: "cimd", which identifies pi with its Client ID Metadata Document on pi.dev instead of dynamic client registration, matching Codex: a shared document when the authorization server sends iss, otherwise a document and callback path specific to the MCP server. Authorization responses are only accepted on the exact redirect URI. Dynamic client registration stays the default.
closes#10302
Move the codemode script reference, including the models API, into
docs/codemode.md and keep only one line per global in the tool
description. Declared tools get a one-line note on how scripts call them,
and the codemode system prompt guidance and MCP server section are shorter.
With the default tools a GPT-5.6 request drops from about 5,300 to 3,300
tokens.
Errors now point the way back: unknown tools and models members name close
matches, models.classify() and models.generateImages() validate their
arguments, unknown or mistyped models point to getAvailableOfType(),
store() overflow explains the store, and unshown generated images get a note.
Scripts run image models with the session's credentials, like
models.classify(). Results are base64 image blocks that image() attaches
to the codemode result; usage counts toward the session cost.
The top-level /login menu offers Sign in with Radius with its status,
with an animated shimmer on the Radius row and a short Radius intro.
After a Radius sign-in it offers to configure the Radius MCP server in
the global mcp.json with auth.provider and reloads. Cancelling a login
returns to the menu it was started from. OAuth sign-ins without a
subscription are labeled as accounts.
Co-authored-by: niloc <5925270+colindaymond@users.noreply.github.com>
The session restored its tool loadout before MCP servers reconnected, so deferred tools that tool_search had loaded were dropped. Restored tools that are not registered yet now stay pending and are activated when they register, until a setActiveTools() call deactivates a tool or the next prompt starts.
The configured metadata document replaces discovery and is trusted as
configured. Authorization codes are only exchanged when the response's
iss parameter names the flow's authorization server.
closes#10172
HTTP MCP servers can set "auth": { "provider": "<provider>" } to send the provider's current /login token as the bearer token instead of using MCP OAuth. The token is read on every request, so provider refreshes apply and nothing is copied to mcp-auth.json. Only allowed in the global mcp.json and from extensions, and requires https except on loopback hosts.
MCP tool and namespace names now replace - with _ like Codex, so the
pi tool name is also the codemode identifier. Tools of a server whose
names collide all get a hash suffix, and server names that differ only
in - and _ are rejected.
closes#10239
When the session's initial tools come from defaultTools, /reload activates
names added to the resolved selection. Removals do not deactivate tools, and
explicit --tools/--no-tools/--no-builtin-tools keep overriding the setting.
closes#10245
In codemode.mode "only", tools whose declarations codemode hides were still
listed in the <tools> section. Drop their snippets so the list matches the
declared tools.
closes#10192
The first prompt waits only for servers with direct tools. Other servers
connect in the background; codemode scripts, tool_search, and the
resource tools wait for the servers they need. codemode and tool_search
are activated from the config.
Tool declarations no longer depend on connected servers: the codemode
description leaves out deferred tools, and the tool_search description
no longer lists sources. Servers are listed in an mcp_servers system
prompt section instead, set in before_agent_start from the config and
connected state, so changes are appended to the conversation.
describeNamespace() and searchTools() accept mcp__dev-radius,
mcp__dev_radius, dev-radius, and dev_radius.
closes#10212
Reject extension commands without a non-empty string name or a handler
at registration, instead of letting them crash the editor when typing /.
fixes#10054
getBranchSelection looked up the model catalog once per assistant message,
which made prompt submission scale with session length. Walk the branch
backward instead: only the last model_change can hold against later
responses, so at most that model needs a lookup.
closes#10198
Adds oauth.clientName (and pi mcp add --oauth-client-name) to set the
client_name sent during dynamic client registration. Defaults to pi.
closes#10226
