You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
SageOx is agentic context infrastructure — we call it the hivemind. It makes architectural and product intent persistent and automatically available across humans and agents.
6
+
**The hivemind for human–agent teams.** SageOx makes your team's decisions,
7
+
conventions, and architectural intent persistent — and loads them automatically
8
+
into every coding session, human or AI.
7
9
8
-
This initial version is intended for AI-native teams — teams that build products almost exclusively through prompts.
10
+
Today, that context lives in scattered places: a meeting nobody recorded, a Slack
11
+
thread three weeks deep, the head of whoever wrote the code. So humans repeat
12
+
themselves and AI coworkers drift — rebuilding the same lost context every
13
+
session. `ox` closes that gap. Decisions become shared memory; every session
14
+
starts with the full picture instead of from zero.
9
15
10
-
Sessions, ledgers, and team knowledge ensure your AI coworkers understand your project's patterns, security requirements, and architectural decisions from the start, making agentic engineering multiplayer by default.
16
+
**Ask your coding agent what the team already figured out — even if it happened
17
+
in a different agent, on a different machine, days ago.**
11
18
12
-
## Demo
19
+
13
20
14
-
21
+
One session, every moment that matters: it **recalls** a teammate's Codex work
22
+
and your own prior Claude Code session, **coordinates** by murmuring and noticing
23
+
a teammate already in the same files, and proposes a plan **enriched** with
24
+
collisions, prior art, and who owns the area. All of it because every session is
25
+
recorded to a shared, queryable **Ledger**.
15
26
16
-
## Install the CLI
27
+
## The loop
28
+
29
+
```mermaid
30
+
flowchart LR
31
+
REC["Record at sageox.ai<br/>discussions, decisions"]
32
+
KB["Knowledge Bubbles<br/>team context plus ledger"]
33
+
PRIME["ox agent prime<br/>loads context"]
34
+
WORK["AI Coworker session"]
35
+
CAP["Session captured<br/>to the Ledger"]
36
+
REC --> KB --> PRIME --> WORK --> CAP --> KB
37
+
```
38
+
39
+
Record a discussion once. It's distilled into **Knowledge Bubbles** — your team's
40
+
shared memory. The next time anyone primes a coding session, that context flows
41
+
in. The session itself is captured back to the **Ledger**, so the next coworker
42
+
inherits it too. Context compounds instead of evaporating.
43
+
44
+
## Works with your coding agent
45
+
46
+
`ox` is agent-agnostic. The same recorded context is primed into, and queryable
47
+
from, whichever agent your team uses.
48
+
49
+
| Agent | Prime context | Record sessions | Murmur (coordinate) | Query past work |
✅ supported · ➖ planned. Claude Code is the primary, most-tested target.
57
+
58
+
<sub>Agent names are trademarks of their respective owners; `ox` is compatible
59
+
with, not affiliated with, them.</sub>
60
+
61
+
## Install
17
62
18
63
**Quick install (macOS / Linux / FreeBSD):**
19
64
@@ -28,91 +73,64 @@ git clone https://github.com/sageox/ox.git && cd ox
28
73
make build && make install
29
74
```
30
75
31
-
## Set up ox in your repo
32
-
33
-
```bash
34
-
# cd into your code repo (e.g. ~/src/my-project)
35
-
cd~/src/my-project
36
-
ox login
37
-
38
-
# one time setup, done ONCE per repo
39
-
ox init
40
-
# commit the changes in your repo, e.g. git commit -a -m 'SageOx init'
41
-
42
-
ox doctor
43
-
# ox doctor --fix may be needed in the alpha stage
44
-
ox status
45
-
# will give you the location of the team context and ledger repos
46
-
```
47
-
48
-
## 👥 Go to [sageox.ai](https://sageox.ai) - Setup team
49
-
50
-
Go into your newly created team in SageOx and invite your coworkers by copying the invite in the upper right, displayed in Team Overview.
51
-
52
-
<!-- TODO: add screenshot of team invite UI -->
53
-
54
-
## 🎙️ Record discussions
55
-
56
-
Team discussions impacting the product are captured and transcribed in the app and the context is automatically available to Claude.
57
-
58
-
<!-- TODO: add screenshot of transcription UI -->
76
+
Verify with `ox version`.
59
77
60
-
## 🤖 Capture sessions
61
-
62
-
`ox-session` capture the conversation between a developer and Claude so the decisions, patterns, and reasoning become available to the rest of the team.
| Load an expert AI coworker into context |`ox coworker load <name>`| — |
113
+
| Diagnose and auto-fix your setup |`ox doctor --fix`|[doctor](docs/reference/doctor.mdx)|
91
114
92
-
Here's a real example from [PR #4](https://github.com/sageox/ox/pull/4):
115
+
## How it works
93
116
94
-
```mermaid
95
-
flowchart LR
96
-
subgraph discuss ["🎙️ Ryan and Ajit Discuss"]
97
-
A1["Captured in Team Context"]
98
-
end
99
-
subgraph implement ["🤖 Ajit Implements"]
100
-
B1["Captured in Ledger"]
101
-
end
102
-
subgraph review ["🚀 Ryan Reviews"]
103
-
C1["Code + Session + Transcript"]
104
-
end
105
-
discuss --> implement --> review
106
-
```
117
+
`ox init` writes a `.sageox/` directory that ties the repo to your team. From
118
+
then on, `ox agent prime` injects team context — conventions, security
119
+
requirements, architectural decisions, prior sessions — into each AI coworker
120
+
before it writes a line. Coworkers, human and AI, share that context through the
121
+
**Team Context** and the per-repo **Ledger**.
107
122
108
-
Ryan and Ajit discussed the daemon design in a recorded team discussion. Ajit then implemented it with Claude (session captured in the ledger). When Ryan reviewed the PR, he had the full picture — the original discussion audio, the implementation session, and the code.
123
+
The payoff is multiplayer by default. When a discussion, an implementation
124
+
session, and the resulting code all carry the same shared context, a reviewer
125
+
opening the PR has the full story — the original reasoning, the session that
126
+
built it, and the diff — without chasing anyone down.
Copy file name to clipboardExpand all lines: cmd/ox/release_notes.md
+18Lines changed: 18 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,6 +7,24 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
7
7
8
8
## [Unreleased]
9
9
10
+
## [0.10.0] - 2026-06-10
11
+
12
+
### Added
13
+
14
+
-**Cross-agent HTML plans with a human review loop** — `ox plan` renders a team-context-enriched plan as a self-contained HTML page for *any* coding agent, not just Claude Code, backed by a catalog of reusable visualizations (timelines, collision maps, sequence diagrams). A new review loop (`ox plan review`) serves the plan for inline human feedback, so a reviewer can shape the approach *before* any code is written — and the feedback is captured back to the ledger.
15
+
16
+
### Changed
17
+
18
+
-**Natively aligned Knowledge Bubbles in `ox status`** — the bubble listing now uses a label-column layout with owner and bubble names aligned in their own column, so the section reads as a clean table rather than ragged, variable-width rows.
19
+
20
+
### Security
21
+
22
+
-**Adapter integrity, a single redaction chokepoint, and environment + clone hardening (ADR-022)** — installed AI adapters are integrity-verified before use, every redaction path now funnels through one chokepoint that is far harder to bypass, and both environment-variable handling and team-context clones were hardened against credential leakage and untrusted-prompt injection.
23
+
24
+
### Fixed
25
+
26
+
-**`ox agent tasks` no longer panics when called with no subcommand** — invoking `tasks` bare previously sliced an empty argument list and crashed; it now guards the slice and prints usage.
0 commit comments