Install and authenticate OpenCode on the machine running your environment, then enable it in Settings > Providers. See provider setup. T3 Code requires OpenCode 1.14.19 or newer, including when you connect an existing OpenCode server.
T3 Code supports OpenCode 2.0.18 and newer. It detects the version on its own, so the same provider settings work for OpenCode 1.x and 2.x. OpenCode 1.x shows Limited support in its provider settings.
OpenCode 2 is a separate package, @opencode/cli. To move from 1.x, install it
yourself, for example npm install -g @opencode/cli. Then refresh provider status.
T3 Code's update button updates whichever package you have installed. It never
switches a 1.x install to 2.x.
OpenCode 2 converts the shared OpenCode database to its own format the first time it runs. Don't run OpenCode 1.x and 2.x side by side on the same machine. Threads you started on 1.x continue on 2.x.
OpenCode compacts long conversations itself, and T3 Code has no setting for it. On
OpenCode 2, change it under compaction in an OpenCode configuration file, such as
opencode.json: auto turns
automatic compaction on or off, buffer sets the token headroom, and keep.tokens
sets how much recent context stays. OpenCode 1.x configures compaction differently.
Send /compact to compact a thread yourself.
OpenCode 2's native subagents inherit the parent thread's allowed T3 tools, but not
T3's per-session tool guidance. T3 directs agents to use delegate_task for child
work that needs project memory, conversation summaries, or code graph. Those
delegated threads receive T3's tools and instructions before they start work.
Leave Server URL empty to let T3 Code start OpenCode locally. A password in
provider settings applies to both that server and T3 Code's connection. With no
password setting, the local server uses OPENCODE_SERVER_PASSWORD from its
environment.
To use an existing OpenCode server, set Server URL and its password in provider
settings. T3 Code uses only that configured password for an external server; it
does not forward a local OPENCODE_SERVER_PASSWORD. If connection or version checks
fail, check the URL, credentials, and OpenCode version, then refresh provider status.
Configured servers, including localhost URLs, do not receive T3 Code's agent tools. Project memory, conversation summaries, code graph, delegation, and other T3 tools are unavailable to the agent in this mode. Leave Server URL empty to use these integrations. This applies to both OpenCode 1.x and 2.x.
When Server URL points to a locally running server (such as 127.0.0.1),
the update button upgrades its installed CLI. If that server is this machine's
OpenCode background service, T3 Code then restarts it on the new version. The URL
must use the service's own protocol, host, and port. For a service listening on
0.0.0.0 use 127.0.0.1, and for [::] use [::1].
Restarting interrupts running turns, and the service then runs with T3 Code's
environment. Set any variables it needs with opencode service set env <name> <value>.
T3 Code does not restart any other server, including the background service
reached through localhost. That server can still report the previous version
until you restart it. For a loopback URL, T3 Code shows that restart requirement
instead of treating the CLI update as a failure.
T3 Code waits for a local server it started to recover after a lost connection. You can stop the turn while it reconnects. If the server exits or an external connection cannot be restored, send another prompt to resume the same OpenCode session.
OpenCode handles retries while a turn is running. If its retries run out on a rate limit and the provider reports when to try again, choose Resume at reset on the failed thread. T3 will continue that thread at or after the reported time. A service overload can also return HTTP 429. OpenCode handles those retries itself.
OpenCode can use the accounts on an inference host, next to its own logins. In Settings > Inference accounts, choose OpenCode beside the host, or turn on Inference routing in OpenCode's provider settings. The host's models then appear in the model picker under T3 inference.
OpenCode and Pi can run every model the host has an account for. On the host, Add an account also offers Antigravity, xAI, Kimi, Meta AI, and Devin sign-ins. Claude and Codex each use only their own accounts.
T3 Code must start the OpenCode server itself, so this is unavailable with a Server URL. Choose the route again in Settings to refresh the model list after changing accounts.
OpenCode follows the shared permission modes. Auto has
the same rules as Supervised because OpenCode has no AI approval reviewer.
Environment files such as .env and .env.local need approval in restricted
modes even though normal file reads do not; .env.example is allowed.
Allow for workspace applies to matching requests in other OpenCode sessions using the same workspace. It is broader than the current thread, especially on a shared external server. Use Allow once for a single request. Denying an action does not stop the whole turn.
After changing an OpenCode login or configuration, use Refresh provider status in Settings > Providers for that environment. On mobile, use Refresh models in the thread settings. Reconnecting also refreshes the catalog; periodic provider health checks do not.
Credential changes are read on refresh. Native OpenCode configuration can remain cached while the local helper is running. Let it sit for 30 seconds without model refreshes or text-generation work, then refresh again to reload the files. Repeated refreshes keep the helper alive. An external server may need its own reload or restart before T3 Code can see configuration changes.
Models defined in a project's own opencode.json appear only in that project's
model picker, after T3 Code reads the project's catalog the first time the picker
opens there.
Existing threads keep their selected model and options even when it disappears from the catalog. If OpenCode rejects that model, select an available one and retry.