Reads your disk
The transcripts your agents already write. Read-only, always.
Your coding agents. One live view.
Run more agents / Lose less context
- Read-only
- No account
- Local by default
- Open source · MIT
Know what needs you
Active, waiting, idle, or inactive across every
repo and branch.
Keep the thread
Goals, decisions, files, commits, and every
command the agent ran.
Keep going
Resume with context intact, or hand a brief to
a fresh agent.
CC
Claude CodeLive now
CX
Codex / ChatGPTLive now
CU
CursorLive now
OC
OpenCodeLive now
PI
PiLive now
GK
Grok BuildLive now
+
More agentsComing soon
01 —
Sessions
Keep moving without losing context.
One place to see what your coding agents are doing.
Waiting on youquestions, permissions, failures
Actively workingwith subagent lineage
Worth a lookwindow filling, cache going cold
coSlashRun more agents. Lose less
context.Settings
⌕ Search sessions --
title, repo, prompt, recap
10 sessions
active in this scope
· 12M tokens
· $19.62 at
list API pricesThis week Board
Table
SESSIONWHERESTATUS /
CONTEXTUPDATED≈ COST
⌄ Active
2
Migrate
settings to the
new token
systemThreading the
theme through
the settings
tree.
acme-dashboard ·
feat/design-tokensCodex
● Active54% context ·
warm
cache
just now
$9.329.9M
tokens
Profile the
batch loader on
A100Tracing the GPU
saturation in
the first
pass.
acme-ml ·
perf/tokenizer-benchClaude
Code
● Active31% context ·
warm
cache
2m ago
$4.103.7M
tokens
⌄ Waiting
1
Fix
double-charge on
retried
checkoutAsked where the
idempotency key
belongs.
acme-payments ·
fix/double-chargeCursor
● Waiting38% context ·
warm
cache
40m ago
$2.482.1M
tokens
⌄ Idle
3
Sweep the eval
harness for
flaky seedsThree seeds
fail under the
fused kernel
path.
acme-ml ·
exp/seed-sweepOpenCode
● Idle44% context ·
cold
cache
1h ago
$1.861.9M
tokens
02 — How
it works
01
02
03
Your work stays on your machine.
Builds the record
Goals, decisions, files, commands, and subagent lineage, indexed locally.
Puts you back in
Resume in your own CLI with full context, or handoff a brief into a fresh session.
coSlash skills for Claude Code and
Codex
Ask your agent to find sessions, make a handoff, send work to another agent, review a session, or diagnose sources.
03 —
Remote host
Local or remote. Same list.
Add a host from your SSH config to see its sessions alongside local work.
Read-only. Asks before installing.
How it stays read-onlycoSlash
— Sessions
127.0.0.1:8787
⌕ Search sessions
-- title, repo, prompt,
recap
2 sessions
on SSH
workspaceThis week
Board
Table
SESSIONWHERESTATUS /
CONTEXTUPDATED≈ COST
⌄ Active
1
Profile the
batch loader
on A100Tracing GPU
saturation
in the first
pass.
gpu-01 ·
acme-mlClaude
Code
●
Active31% context
· warm
cache
just now
$4.103.7M
tokens
⌄ Idle
1
Sweep the
eval harness
for flaky
seedsThree seeds
fail under
the fused
kernel
path.
gpu-01 ·
acme-mlCodex
●
Idle44% context
· cold
cache
5h ago
$1.861.9M
tokens
04 —
Resume or hand off
Resume or start fresh.
Check context, cache, and the debrief before you choose.
91% context, compacted 2×, 14 behind mainthis one says start fresh
05 —
Pricing
Free locally. Pay for collaboration.
Individual
Available now
$0forever
For individual developers
- Local attention board
- Claude Code, Codex, Cursor, OpenCode, Pi, Grok Build, and subagents
- Optional read-only SSH monitoring of one Linux host
- Title, repository, and branch search
- Readiness signals, resume, and handoff
Team
In development
Coming soon
For shared agent work
- Shareable live visibility
- Shareable session discovery
- Cross-person handoffs with context
- Workspace access controls
- SOC 2 Type II
Enterprise
Custom
For organizations with advanced needs
- Everything in Team
- Enterprise deployment options
- Security and compliance support
- Priority onboarding and support
06 —
FAQ
The important questions, answered.
What does coSlash read?
The session transcripts Claude Code, Codex, Cursor, OpenCode, Pi, and Grok Build already write to disk, plus the Git and process metadata needed to tell a running session from a finished one. There is nothing to connect and no account to create — it reads what is already on your machine, plus one Linux host over read-only SSH if you choose to add one.
Which agents does it support, and on what?
Claude Code, Codex, Cursor, and OpenCode, including local desktop app and CLI sessions, plus Pi and Grok Build CLI sessions. Cursor IDE chats can reopen their workspace; Cursor CLI sessions can resume exactly. coSlash runs on macOS (Apple Silicon or Intel) and Windows 11 amd64. Linux sessions can be monitored over SSH — see the next answer.
Can it see agents running on a remote server?
Yes — one Linux host, and only once you
add it. In Settings → Machines you
give coSlash an alias that is already in your
OpenSSH config; it uses the system
ssh client and never edits that
config. It checks SSH and SFTP, then asks before
installing a small collector in the SSH
user’s own home directory — no root,
no network access, and a fixed allowlist of
agent transcript paths. Read-only SFTP is the
fallback if the collector cannot run.
Remote sessions join the same table as local ones with their host, status, costs, tokens, and file-edit summaries. Resume and handoff work while the host is connected. When remote process state cannot be confirmed, coSlash labels its liveness as unknown. Synthesis and commands stay local-only.
Does coSlash change my code or my transcripts?
No. Transcripts are opened read-only and never
written to, and coSlash makes no commits,
branches, or edits in your repository. It writes
only to its own directory,
~/.coslash, where cached debriefs
and pending handoffs live. Resume and handoff
launch your own CLI in your own terminal —
every edit after that is the agent’s,
exactly as if you had started it yourself.
Does anything leave my machine?
Not unless you turn it on. Debriefs are derived from the transcript with no model involved. Optional AI synthesis only sharpens them, and stays off until you enable and save it. When on, it sends a bounded set of derived facts — about 12 KB — through the local agent CLI you pick. That spends a little of the agent’s own usage: a few cents per session at most, and only substantial sessions qualify.
It runs a local server — is that exposed?
No. coSlash binds to
127.0.0.1:8787 and nothing else,
and every API request carries an access token
minted when the server starts. It is not
reachable from your network, and a link from an
earlier run stops working once the server
restarts.
How are costs calculated?
They are estimates at list API prices, not invoices: token counts from the transcript against a dated per-model price table. A model with no verified price is flagged and left out of the total rather than guessed at, so the number is never quietly wrong. OpenCode sessions report the cost they recorded themselves instead of an estimate.
Can I pick a session up in a different agent?
Yes. Resume reopens the original session in its own CLI with full context. Start fresh writes a Markdown handoff — objective, current state, key decisions, timeline, files, commits, next steps — and opens a new session with it loaded as reference, so the next agent waits for your instruction instead of acting on stale notes. The handoff is plain text, so it crosses vendors: a Claude Code session can be picked up in Codex.
The session ends. The knowledge stays.
Install once, and every session your agents run stays legible — today and three weeks from now.