A llama session was a chat box: no tools, a fixed model, no permission
mode, and a model name drawn as the path the file sits at. It now runs the
agent loop itself, which is what the pieces below all hang off.
Tools are `llama-server`'s own (`--tools all`), which that server both
publishes and runs -- `GET /tools` for the definitions, `POST /tools` to
call one. Web search is Exa's MCP server, reached from this backend rather
than from the machine serving the model: that is what llama.cpp's own web
UI does, and it puts the search on the machine with a route out instead of
the one with the GPU. `llama-server`'s `--mcp-servers-json` can only spawn
local commands, so using it would have meant a Node bridge on every
machine that serves a model.
Driving the loop is what makes the permission gate ours. Two modes,
`manual` and `bypassPermissions`, which is what the mechanism has: the web
UI asks before every call and remembers the tools you say "always" to. The
allowances fold back out of the transcript's own answers, so they survive
a restart and a model change without being stored anywhere else.
Also here, because tools made each of them matter:
- **Loading is a state.** A 12 GB model takes twenty seconds to reach
memory and refuses everything until it has; the session used to report
`running` for that whole time, and a message sent meanwhile came back as
an error. It is `loading` now, and the message waits.
- **The model can be changed.** A `llama-server` holds one model, so this
stops it and starts another. The conversation survives because it was
never in the server.
- **Models are named, not pathed.** `general.name` read out of the file
itself -- over ssh too, in the round trip the spawn was already making.
Where two models share a name the file name breaks the tie.
- **`-np 1`, and the MTP draft head where the file has one.** Measured on
the 27B here: 41.5 tok/s plain, 61.4 with `--spec-type draft-mtp` at one
slot, and 28 with it at four -- speculating against a split KV cache is
worse than not speculating. The flag is conditional because asking for a
head that is not there makes `llama-server` exit.
- **A refusal says what to do.** Tool results are thousands of tokens, so
an overrun context is now ordinary; it was "http status: 400" and is now
the server's own "exceeds the available context size, try increasing it".
`GET /machines/{id}/models` is gone: the provider models route answers the
same question, and two answers to one question is how a picker comes to
offer a model the spawn screen does not.
Verified end to end against real models: a tool call asked and allowed, an
Exa search, a shell command, a 27B loaded while a message waited on it, a
model switch mid-session, a second message queued behind a running turn,
and the whole of it again on a session running over ssh.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A 401 from the usage endpoint means the stored access token has expired.
Refreshing it here is not an option: Anthropic's OAuth rotates the refresh
token, so a second refresher invalidates the CLI's copy and forces a
re-login on a machine that usually has a live session on it. So run the CLI
there instead and re-read what it wrote.
`doctor` rather than `auth status`: probed against 2.1.258 with an invalid
token, `auth status` answers loggedIn:true from the file alone and never
reaches the network. The same probe showed a failed refresh blanks both
tokens, which is why this stays on the 401 path.
Also gives ProviderConfig one program() so the CLI's default path is not
written down twice.
A 401 is the endpoint answering and refusing the stored OAuth token, which
Claude Code refreshes as it runs -- so a machine whose CLI has been idle
hands us a stale one. Reporting it as "usage endpoint unreachable" pointed
at the network instead of at the one thing that fixes it.
The rate-limit bar answered a question about an account, and picked the
answer by machine. One machine runs echo, the Claude CLI and a local
model side by side, so every echo session on it drew the CLI's five-hour
window: a quota that session cannot spend and could never run down. A
session now names its meter (`usageProvider`, from
`DriverKind::usage_provider`, which `usage::providers_for` reads too so
the two lists cannot disagree), and the phone matches on machine *and*
provider. Nothing meters echo or llama, and nothing at all is drawn --
including while the first fetch is out, since "checking" under a session
that turns out to meter nothing is a row the screen then withdraws.
Echo gets a meter it can be *told* about instead: `/usage 42`,
`/usage 95 20`, `/usage 42 never`, `/usage notloggedin`,
`/usage unreachable`, `/usage failed`, `/usage off`. Those states cost
real quota to arrange, which is why none of them had been looked at.
And llama.cpp runs wherever a setup says, which was the last of phase 5.
`Transport::reserve_port` is the second half of what a transport is --
"run this" plus "reach this port" -- returning the port the server binds
there and the port that reaches it here, and `Launch::reaching` puts the
`-L` tunnel on the connection that already carries the command. Three
things that came out of building it:
- A forwarded launch gets a pty and every other one keeps `-T`. Killing
the ssh client ends a CLI by closing the stdin it reads; llama-server
never reads its stdin, so the same kill left it running on the far
machine with the model loaded -- one orphan per stopped session.
- The model is looked for on the machine that will serve it, at that
machine's own models directory, so `GET /setups/{id}/models` is what
the spawn screen offers rather than the backend's own downloads.
- The readiness poll watches the process, not only the port: a model
that will not load exits in a second and would otherwise have been
reported as "gave up after 300s". The failure carries the log's tail.
Exercised end to end against this VM over ssh to itself: spawn, load,
answer, outlive a backend restart, be adopted, answer again, and stop --
with both the ssh client and the far llama-server gone afterwards. The
local path, the Claude bar and the spawn screen checked on the emulator.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The markdown had accumulated a lot that was stale rather than wrong.
PLAN.md still described pi as the llama.cpp harness, a refcounted
LlamaServerManager, and a providers-by-hosts cross-product, all of which
were superseded or never built; it also carried a second copy of the HTTP
table that routes.rs owns. EXPLORER.md and TRANSCRIPT_CACHE.md held
implementation checklists for work that has since landed. AGENTS.md
restated most of PLAN.md's design instead of being the working-notes
layer it says it is. 3225 lines of markdown to 2180, with the stale
sections gone rather than reworded.
On the server, comments explaining what the code already says are out and
the ones recording a constraint, a measurement or an incident are kept but
cut to a few lines each: 5504 comment lines to 4586.
Four doc comments in session/mod.rs, and one each in process.rs and
usage.rs, had drifted onto the item above the one they describe --
functions were reordered without them, so `stop_session`'s doc sat on
`set_session_cwd`, `stat_of`'s on `struct Stat`, and `UsageMonitor`'s on
`type Cached`. Each is back on its own item.
routes.rs's module table also claimed later phases would add `/hosts`,
which setups replaced.
cargo test (127 passed), clippy --all-targets and fmt are clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Uploads no longer sit whole in memory anywhere: the phone writes the
multipart body chunked as it reads the picked file, and the server writes
each chunk to a `.part` file under the session and renames it when whole.
The per-request cap is 4 GB and bounds disk, not memory.
A file attached to a session on another machine is copied there in the
same request: one ssh invocation takes the bytes on stdin into the
setup's `attachmentsDir` (new, optional, on the machine form and in the
config), else the session's cwd, else the login home, and answers with
`pwd -P`, which is recorded beside the file as `<name>.remote` and is the
path the driver tells the CLI. A failed copy fails the upload and says
why, so no message ever names a file that is not there. The host keeps
its copy so transcripts can reference and fetch it. Measured against the
Gentoo test guest: a 40 MB file shared from the phone arrived there byte
for byte. The tilde in that setting is the remote home, so it is not
expanded on the server the way other setup paths are.
TRANSCRIPT_RENDERING.md records the week of transcript work -- the
measurements behind each decision, the harness, what was rejected, and
what to do next -- so a new session can start from it.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A Rust backend that owns the sessions and an Android app that reads them.
The server spawns and adopts CLI processes, normalises everything they emit
into one event model, keeps the transcript, and serves it over pinned TLS on
a WireGuard interface; the phone streams that, replies, sends images, and
imports conversations the machine already has.
`AGENTS.md` is the working guide -- what runs where, what has been measured,
and the faults that were expensive to find. `PLAN.md` is the design record.
History before this point was squashed away. It was a personal project's
running commentary and carried a name and a couple of machine paths that
have no business in a public repository; the tree is what mattered and the
tree is here.
iris-aiandClaude Opus 5
iris