Match @ai-sdk/anthropic exactly: baseURL is the full API prefix, so
/messages is appended as-is instead of unconditionally inserting /v1,
which broke configs where the baseURL already ends in /v1 (the common
form OpenCode itself passes to @ai-sdk/anthropic).
Addresses review feedback from @btriapitsyn on PR #3000.
The Anthropic small-model dispatch (session-assist recap/suggestion,
walkthrough, etc.) hardcoded https://api.anthropic.com/v1/messages,
ignoring provider.anthropic.options.baseURL — the same override every
other provider branch already respects. Any setup routing the anthropic
provider through a local proxy (e.g. Meridian for Claude Max / Vertex)
had the small-model call bypass the proxy and hit the real Anthropic API
with whatever placeholder credential the proxy config uses, failing with
401 invalid x-api-key on every call.
The mobile-vs-desktop surface is stamped once at boot, so a browser
window narrowed past the phone threshold kept the desktop shell (and
its legacy squeezed layout) until a manual reload. A viewport watcher
now reloads into the other shell once the resize settles — the same
mechanism the old Settings toggle used. Fixed shells (Capacitor,
desktop, VS Code) and ?surface= overrides never switch.
With the new mobile app reachable this way, the old/new mobile layout
preference is gone: phones always get the mobile app.
The tab list lived only in per-tab sessionStorage, so a new browser
tab, another device, or cleared storage showed an empty terminal
sidebar while PTYs kept running server-side, and orphans leaked until
the idle sweep. Add GET /api/terminal/sessions and adopt unknown
server sessions into the local tab projection (additive only; a failed
listing changes nothing).
The idle sweep also reaped terminals in background tabs because only
the active tab holds a WebSocket attachment. Add POST
/api/terminal/touch and have open clients periodically refresh
activity for every session their tabs reference.
Two mitigations for local multi-instance contention over the shared relay
identity:
- A standby instance now waits a 2-minute grace period after the host claim
frees before taking over, so a cleanly restarting host (app update or
relaunch) — which reclaims at boot with no wait — always wins the restart
window instead of stranding paired devices on another process.
- Dev instances never host the relay passively: dev scripts set
OPENCHAMBER_RELAY_HOST=off and the Electron dev shell is detected via
OPENCHAMBER_ELECTRON_DEV. Explicit enable/pairing on such an instance still
force-claims; OPENCHAMBER_RELAY_HOST=on overrides.
A dedupe-keyed client re-mint (QR rescan, password/passkey re-login) replaces
the stored record, which also reset the operator-visible name to the app's
hardcoded default ('OpenChamber Mobile'). The app-reported name is now only a
fallback: an explicit pairing label wins, otherwise the replaced record's
label is kept, and the default applies only to a first-ever pairing.
A project could pin the model new chats start on, but not the level to
run it at: the default cascade dropped any variant as soon as a project
model won, and only ever considered the global one — which belongs to
the global model.
Projects now carry `defaultVariant` alongside `defaultModel`, stored and
sanitized only next to that model, and the cascade passes it through.
Both controls sit in one "Defaults for new chats" group laid out like the
Sessions defaults, and the level appears only for models that offer them.
Parakeet is an offline model trained on whole utterances, so re-decoding
the growing buffer to animate a live transcript cost O(n^2) work for a
result the final decode replaced. Sessions now decode once per committed
segment, and the composer shows a scrolling waveform of the mic level
instead of running text.
Long dictations split at a pause once past 60s (hard cap 90s) instead of
on a blind 15s timer, so cuts no longer land mid-word. Committed segments
decode while the user is still speaking: a 185s dictation returns 4.1s
after stop instead of 11.0s, with identical text (816 vs 817 words).
Also fixes two ways the stream manager could silently drop transcribed
audio. It now counts the commits it issued instead of trusting the
session's echoed events, so a commit still in flight when the client
finishes can no longer be left out of the final text. And segment
byte/peak accounting is reset where the commit is issued rather than when
the event arrives, which could mistake the tail of a dictation for
silence and clear it.
Remove retired Command Code, Discord, and Telegram integration entries, search targets, and documentation. Keep Command Code provider usage and logo support available through normalized provider ID aliases.
Show every change on the current branch relative to its base in the
Changed/Staged/Last turn dropdown. The base comes from the branch's
reflog record or an explicit per-branch user choice (persisted), never
a main/master guess; when git has no record the user picks a base once
from a searchable branch list.
- server: GET /api/git/branch-base (reflog-derived base),
GET /api/git/range-files (name-status -z with rename/copy
destination paths and -C copy detection)
- shared UI: optional getBranchBase/getGitRangeFiles runtime APIs
with boundary parsing; persisted per-branch overrides keyed by
runtime+directory+branch
- DiffView: branch scope with confirmed-unavailability coercion of
persisted tabs (detached HEAD, default-branch checkout, metadata
settled without a default), range-invalidated diff cache guarded
against stale completions, bounded branch-metadata retry, read-only
diff actions in branch scope; hidden in VS Code
- helper module branchDiffScope.ts with tests for coercion,
availability, race conditions, and retry exhaustion
* test: add regression test for stat directory scoping (#3019)
* fix: scope file-reference stat probes to the session directory (#3019)
The /api/fs/stat probe sent by the markdown renderer carried no
directory hint, so the server resolved the workspace from
settings.lastDirectory. When the browsed directory differs from the
session directory the stat returns 400 and the renderer treats it as
file-does-not-exist, silently disabling file-reference links.
Send effectiveDirectory as x-opencode-directory on the probe and
qualify the stat cache key by directory so a rejection under one
directory cannot leak into another. The probe logic moves from
MarkdownRendererImpl into fileReferenceStat with unit coverage.
Create projectless chat sessions under a managed, date-scoped Chats directory and clean abandoned or deleted session folders.
Add Chats to sidebar state, startup cache, shared context, and Electron Mini Chat while keeping VS Code project-only. Resolve managed chat directories to one server-side memory owner and document the runtime contracts.
* feat(skills): remove ClawHub catalog integration
Drop the ClawHub registry as a skills catalog source across web server,
shared UI, VS Code, docs, and locales. The catalog now serves git-based
sources only: the curated Anthropic repo and user-defined repositories.
Also removes the now-unused adm-zip dependency.
* feat(skills): redesign catalog around curated GitHub repositories
Replace the single-source dropdown with a card grid of curated GitHub
repositories (Anthropic, OpenAI, Cursor pstack/skills, Matt Pocock) plus
user-defined sources. Source cards show skill counts, GitHub stars, and
last-updated time; a global search covers all loaded sources.
Server: curated sources gain GitHub repo metadata (stars, pushed_at)
fetched best-effort with a 3-hour in-memory and on-disk cache; scans
run through a concurrency-limited, deduplicated cache with 3-hour TTL
persisted across restarts. Refresh still bypasses the cache.
Shared UI: source cards, global search with clear button, per-skill
GitHub links, install/installed states. VS Code curated list updated
to match. All new copy translated across 12 locales.
* fix(skills): address catalog review findings
- GitHub metadata fetch timeout drops to 1.5s (under the catalog
client's 3s deadline) and failed lookups cache briefly (5 min) so
repeated catalog loads do not re-hit a failing API.
- Disk cache files are written with owner-only permissions (0o600);
rename preserves the mode.
- loadSource deduplicates concurrent in-flight requests per source and
the shared isLoadingSource flag now clears only when the last active
source load finishes.
Plugin providers are registered from a plugin's `config` hook and credentialed
from its `auth` loader, both inside the running OpenCode process. Nothing about
them reaches `opencode.json` or `auth.json`, so resolution that only reads files
could not see them: selecting such a model failed with "has no known API base
URL" while the same model worked in chat (#2666).
`GET /provider` is where that state is visible. A new `runtime-providers`
module keeps one cached snapshot of it and reports, per provider, the
credential and endpoint OpenCode itself resolved. Credential resolution becomes
config -> runtime -> auth.json, and endpoint resolution config -> openai default
-> runtime -> models.dev catalog.
Providers with a dedicated wire format (Copilot, ChatGPT-plan OpenAI, Anthropic,
Google) are excluded from the runtime credential: for them OpenCode reports an
OAuth access token that their real transport does not accept.
opencode zen is excluded when the user has no zen login. OpenCode then reports
the sentinel `apiKey: "public"` and trims its catalog to free models that run on
its own infrastructure; the sentinel is never read as a credential.
Claude Code stays refused for background actions even when a plugin publishes an
OpenAI-compatible endpoint for it, because that endpoint is a facade over the
Claude Agent SDK and spawns the CLI per request.
No capability probe. Asking `GET /models` does identify a plugin whose protocol
lives in its own `fetch`, but measured across the 166 providers with an `api`
URL in the models.dev catalog it also denies six that work and simply have no
`/models` route. A provider that vanishes from the picker explains nothing,
while one that fails on use says why, so availability stops at credential and
endpoint.
The same list drives the Small Model and Changes Walkthrough pickers.
Validated against a real OpenCode with four plugin providers loaded: offered
providers went from 3 to 7, zen and Claude Code stayed out, and a generation
through a plugin-backed model that previously failed now returns.
* fix: reconcile busy sessions after managed OpenCode restart
Forced health-check restarts previously rebound the event stream without
settling in-flight turns, so sessions stayed busy with no terminal state.
Interrupt those sessions, classify health failures, and retain bounded
process diagnostics for post-restart diagnosis.
Fixes#2943
Co-authored-by: serkraser <serkraser@gmail.com>
* fix: surface interrupted chats after OpenCode restart
Complete unfinished assistant turns as aborted once the session is
authoritatively idle, and show a persistent toast so users can continue
instead of remaining silently stranded.
Fixes#2943
Co-authored-by: serkraser <serkraser@gmail.com>
* fix: redact Basic auth credentials in restart diagnostics
The key/value sanitizer stopped at whitespace, so Authorization: Basic
credentials survived in stderr tails and health snapshots. Redact the
scheme token before that rule runs.
Co-authored-by: serkraser <serkraser@gmail.com>
* fix(git): create worktrees from forked PRs via refs/pull/<n>/head fallback
A worktree created from a linked GitHub PR whose head branch lives in a fork
failed when the fork's head repository was missing (deleted fork) or
unfetchable (auth, network): the dialog threw 'PR head repository URL is
unavailable' before any git command ran, and the server had no fallback to
refs/pull/<n>/head, which GitHub serves on the base repository.
- NewWorktreeDialog: when pr.headRepo is absent, send a prRef config
(refs/pull/<n>/head from origin) instead of throwing; the fork config now
also carries prRef so the server can fall back when the fork fetch fails.
- git service: fetchPullRequestHeadRef fetches refs/pull/<n>/head into
refs/remotes/<remote>/pr-<n>-head (same refspec shape as
fetchRemoteBranchRef) and both validateWorktreeCreate and
attachGitWorktreeToCandidate fall back to it when the fork path fails;
fallback worktrees get --no-track and no upstream config because a PR ref
is not pushable. When both paths fail the original fork error surfaces.
- Focused tests cover the prRef-only path and the fork-unreachable fallback.
Fixes#2422
* fix(git): harden PR worktree fallback against stale fork refs (#12)
After a fork fetch fails, resolve immediately from refs/pull/<n>/head
instead of accepting a cached remotes/<fork>/<branch> tracking ref.
Match the PR base repository by URL (not a hardcoded origin remote),
store fetched PR heads under refs/openchamber/pull/<n>/head, and share
one existing-mode resolver between validate and create.
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* fix(git): make PR head SHA authoritative and namespace private refs
Reuse local/remote branches for linked PRs only when their tip matches
pr.headSha; otherwise fall through to fork fetch / refs/pull. Store PR
heads under refs/openchamber/github/<owner>/<repo>/pull/<n>/head, prefer
HTTPS for direct base-repo fallback, and surface composite fork+fallback
errors when both paths fail.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* test(ui): assert validate/create forward deleted-fork PR payload fields
Guards the dialog wiring regression where validate omitted prRef while
create included it, by asserting worktreeManager forwards prRef,
prBaseRepoUrl, and related fields for deleted-fork configs.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* refactor(git): always checkout linked PRs from refs/pull/<n>/head
Move PR worktree resolution to the server. The UI now sends only
pullRequest identity (number + baseRepoUrl + optional head fields);
the server always fetches the authoritative PR head and best-effort
configures fork upstream afterward.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* refactor(ui): drop PrWorktreeConfig; send PR identity only
Delete the prWorktreeConfig module. NewWorktreeDialog maps linked PRs
straight to pullRequest identity, skips upstream defaults for that path,
and leaves checkout + optional fork tracking to the server.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* refactor(git): linked PRs are {number, baseRepoUrl} only
Drop fork upstream / tracking and head/base owner-repo fields from the
linked-PR worktree path. Fetch refs/pull/<n>/head, create --no-track, done.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* fix(git): meet #2422 Must/Should without refs/pull fallback
Linked PRs send fork identity only; the server provisions pr-<owner>,
fetches the head branch, and fails clearly when the fork is missing or
unreachable. Local reuse requires a matching headSha. Prefer HTTPS for
headRepoUrl. Do not write upstream tracking when the upstream ref was
never fetched.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* fix(git): drop invalid upstream fallback and PR branch collisions
Remove setBranchTrackingFallback: if upstream fetch fails, leave tracking
unset. When a linked PR's head branch already exists locally with a
different tip, create pr-<number> instead of git worktree add -b on the
colliding name.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
* fix(git): strip PR worktree create back to fork-remote provision (#15)
Keep the original ensureRemoteName/Url path for linked fork PRs, prefer
HTTPS clone URLs, fail clearly when the fork is unreachable, and leave
upstream tracking unset when the upstream ref was never fetched.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
The panel stored notes, todos and plans inside one shared JSON file that
six unrelated domains also wrote to, synchronised itself through window
CustomEvents, and could only read plans. It is now Project knowledge:
server-owned storage with explicit routes, a store with rollback, a
section sidebar, plans that open and edit in place, and search across
all of it.
Notes and plans the user pins travel with every message sent in that
project. Pinning is project state, not an attachment to one message, so
it holds until unpinned and the work status panel names what is riding
along and can detach it.
Agent memory is added alongside, in two scopes: what is true about the
user, and what is true about this codebase. The split is not cosmetic —
a wrong project fact costs one project and is noticed, while a wrong
global fact quietly shapes every session everywhere and the user has no
code to check it against. It stays separate from notes so an agent
mistake cannot land in what the user wrote. Sessions receive an index of
titles only; bodies are read on demand, because an index carrying full
text grows until it crowds out the conversation.
Deciding what a session must be told, and whether it has been told, now
lives on the server. The client owned it before, which meant sessions
started without a UI — scheduled tasks, sessions the agent dispatches —
received nothing at all, and a tab's record of what it had sent outlived
the conversation: after compaction the agent no longer held the block
while the tab went on believing it did. What was delivered is recorded
in the session's own metadata, and compaction restores it through the
runtime that already restores pinned messages, in the same turn.
Agent memory ships dark behind OPENCHAMBER_MEMORY_ENABLE: unset, there
is no tool, no routes, no session index, no settings row and no panel
tab. Absent rather than switched off, so nothing invites turning on a
feature that has not been announced. Pinned notes and plans are
unaffected and ship as normal.
* fix(proxy): reuse upstream connections for OpenCode API requests
`createProxyMiddleware` was constructed without an `agent`, so `http-proxy`
fell back to `agent: false`. That disables connection pooling and forces
`Connection: close` on every proxied request, consuming one ephemeral port
per request.
Measured against a real `opencode serve` instance, 200 sequential requests
through the proxy created 201 TIME_WAIT entries (1.005 ports/request). With
a keep-alive agent the same load creates 0.
On macOS the ephemeral range is 16,384 ports and TIME_WAIT lasts 30s, so
sustained traffic around 546 req/sec exhausts the pool — after which every
process on the host fails to open outbound connections with EADDRNOTAVAIL.
`maxSockets: Infinity` preserves the unbounded concurrency of `agent: false`,
so this changes connection reuse only, not request throughput.
Partially addresses #2915.
* fix(proxy): derive proxy agent class from the target scheme
Addresses review feedback on #2916. The first commit created an
unconditional `http.Agent`, which regresses external OpenCode servers
configured over https via `OPENCODE_HOST` (accepted by env-config.js).
http-proxy dispatches through `https.request` when the target protocol is
`https:` (http-proxy/lib/http-proxy/passes/web-incoming.js:126), and
`http.Agent#createConnection` is plain `net.createConnection` — so an
http.Agent would open a plaintext socket to a TLS port and fail every
proxied request. `agent: false` previously worked for both schemes.
`createOpenCodeProxyAgent(target)` now returns an `https.Agent` for https
targets and an `http.Agent` otherwise, derived once from
`resolveProxyTarget()` at registration so the single shared instance is
preserved across `apiProxy` and `interactiveOAuthProxy`.
Guarded in both test layers, verified to fail when the selection is
reverted to an unconditional http.Agent. `https.Agent` extends
`http.Agent`, so the http cases assert `not.toBeInstanceOf(https.Agent)`.
* Round 2: fix: resolve the proxy agent lazily so cold starts honor https
Addresses the round-2 blocker on #2916. Deriving the agent class at
registration is too early: startup-pipeline-runtime.js calls setupProxy()
(line 104) before bootstrapOpenCodeAtStartup() (line 141), so on a fresh
process state.openCodePort is null, buildOpenCodeUrl() throws
(network-runtime.js:86-88), and resolveProxyTarget() returns the http
loopback fallback. An external server configured via OPENCODE_HOST=https://
only appears on state.openCodeBaseUrl after bootstrap, so it was still
getting a plain http.Agent — the regression the previous commit intended
to fix.
`agent` is now a getter backed by a per-scheme memoizing resolver.
http-proxy-middleware rebuilds per-request options with
`Object.assign({}, this.proxyOptions)` in prepareProxyRequest, which invokes
getters, so resolution happens at request time while still yielding one
shared pool per scheme.
Tests now model the production ordering — registration while the port is
null and buildOpenCodeUrl throws, then an https base URL appearing after
bootstrap — and fail against the eager implementation. A behavioral test
pins the http-proxy-middleware option re-read the fix depends on, so a
library change that froze options would fail loudly instead of silently
regressing https targets.
The resolver is module-private; `bun run dead-code` flagged it as an
unused export when it was exported.
* Round 3: docs(changelog): note upstream connection reuse under [Unreleased]
Repo precedent adds [Unreleased] bullets for comparable proxy/stability
fixes (1.18.4 Stability, 1.9.3 Reliability/Proxy). Non-blocker raised in
review on #2916.
* Round 3: docs(changelog): use repo-standard 'behavior' spelling
* Round 4: docs(changelog): don't imply a restart is the only recovery
The ephemeral port pool drains on its own once the exhausting traffic
stops (TIME_WAIT expiry), so a restart is sufficient but not necessary.
Optional nit raised in review on #2916.
* Round 5: fix: construct the proxy agent through one factory; widen the pool
Review found the https branch was mutation-uncovered: the resolver
re-implemented agent construction inline instead of calling the exported
`createOpenCodeProxyAgent(target)`, so replacing its https branch with
`new https.Agent()` — dropping OPENCODE_AGENT_OPTIONS, and with it
keep-alive — left the entire suite green. Since `createOpenCodeProxyAgent`
also had no production callers, its four tests were pinning dead code.
Delegating collapses both: the factory is now the single construction
path, and the mutation fails 2 tests including the live resolver path.
Also from review:
- maxFreeSockets 32 -> 256 (Node's own default). The lower cap evicted
pooled sockets under concurrency, reintroducing the churn this agent
exists to prevent: at 64 concurrent requests it left 303 sockets in
TIME_WAIT versus 0 at 256.
- Added `timeout` to OPENCODE_AGENT_OPTIONS. Free-socket eviction is
governed by agent.options.timeout, which was unset, so idle sockets
persisted until the peer closed them. `keepAliveMsecs` is the TCP probe
delay, not the idle lifetime.
- resolveProxyTarget() now checks openCodePort before calling
buildOpenCodeUrl instead of relying on it throwing. The port is nulled
on several runtime paths (health-check failure, failed restart), so a
degraded OpenCode made every proxied request pay for a thrown-and-caught
exception — and the getter added a second call per request.
- Test fixtures use :4096 rather than :443; WHATWG URL elides the default
port, so parseInt('') is NaN and env-config rejects that host. The
fixtures modeled a state that cannot reach production.
- The getter-read assertion is now exact (0 at construction, 1, then 2)
rather than >= 2, which would have passed if the getter were read twice
at construction and never per-request.
- listen() rejects on 'error' and servers start inside try/finally, so a
bind failure fails the test instead of hanging to timeout.
Treating every undefined parse as empty config let a file that is not JSON
at all (YAML, plain text) read as {}, so a later write would back it up and
replace it - the same data loss this fix is meant to prevent. Only a
comment-only parse, where ValueExpected is the sole error, counts as empty.
Branch status resolves an open PR across the whole fork network first, so a
merged fork PR can never hide an open upstream PR for the same head. Only when
no target has an open PR does the branch's newest closed/merged PR come back,
as history.
The panel shows that history as a compact note and offers creating the next PR
below it, instead of either sticking on a terminal PR or going blank after a
merge. Terminal associations stay persisted for reload continuity but are never
treated as authority: they revalidate on the discovery cadence and on focus.
History is looked up only for the branch's own remote and name, and remembered
per repo+branch, so the extra lookup cannot exhaust the route's resolve budget.
The checks summary and merge-permission lookup are skipped for a closed or
merged PR, where neither is actionable.
Auto-derived project labels were title-cased, turning .ssh into .Ssh and
opencode-claude into Opencode Claude. Show the folder name verbatim in the
sidebar, window title, settings selector and notification templates, and
migrate persisted legacy labels back to the folder name (manual renames
are preserved).
Plugin listing still called readConfigFile per layer, so one unparseable
project file made GET /api/config/plugins and VS Code listPluginEntries
fail. Reuse readConfigLayer isolation and pin comment-only empty parse
in the VS Code suite.
Co-authored-by: serkraser <serkraser@gmail.com>
Keep both Unreleased bullets: the config-wipe Settings fix and the
stale merged-PR Git panel fix from main.
Co-authored-by: serkraser <serkraser@gmail.com>
Address review findings on the #2923 fail-closed parse fix:
- Comment-only files produce ValueExpected with no JSON value; treat that
as empty config instead of INVALID_JSONC. Partial object trees still throw.
- readConfigLayers no longer lets one unparseable layer abort valid sibling
layers. Mutations still fail closed on the custom/user write target.
Co-authored-by: serkraser <serkraser@gmail.com>
Add store persist/hydrate regressions for terminal branch associations
and a server test that a complete empty open list does not query closed
PRs. Inline the open-only matcher state so the next repo target can win.
Co-authored-by: serkraser <serkraser@gmail.com>
Branch PR status now resolves open PRs only, revalidates closed/merged
associations on a discovery cadence, and clears authoritative empty
results so the panel can self-heal without a manual refresh.
Co-authored-by: Serhii Dziupin <makeittech@users.noreply.github.com>
jsonc-parser was returning truncated trees (often only `$schema`) when
configs contained JSON5-style unquoted keys. Config mutations then backed
up and overwrote the full file with that stub. Check parse errors on read
and refuse to overwrite unparseable files on write in web and VS Code.
Fixes#2923
Co-authored-by: serkraser <serkraser@gmail.com>
Matt Peter