Arkon Vault — private project memory so work continues across ChatGPT, Claude, and Codex
What we shipped
Arkon Vault is a continuity layer between the AI tools you already use — not another assistant.
One private workspace holds project vision, state, decisions, and handoffs. Connected AIs preflight the authoritative next step, work against that project, save a handoff when you pause, and continue later (including in a different AI) without rebuilding the brief.
How you connect
- ChatGPT: Plugins listing → Install → approve Vault workspace → @Arkon Vault → preflight
- Claude: Community connector → connect → approve workspace → ask Claude to run preflight
Normal setup is marketplace/OAuth — not a pasted server URL.
Pricing (honest catalog)
- Solo $3/mo
- Team $9/mo
- Ops $22/mo
Billing page is authoritative. Cancel via the verified portal. No launch discounts or free months in this post.
Privacy
Deliberate Vault tool input only — no silent capture of every chat. Per-AI revoke, export, and deletion controls. Optional Windows Sync does not upload complete raw transcripts.
https://vault.arkoncybersecurity.com/privacy
Optional design-partner path
First Five: five qualified continuity tests for people who use 2+ AIs — content-safe feedback only.
https://arkoncybersecurity.com/first-five
Links
Product: https://arkoncybersecurity.com/
Support: support@arkoncybersecurity.com
Work should continue.
— Arkon Cybersecurity
Preflight is the right instinct — it turns 'continuity' from passive memory into a coordination mechanism, which is the part most memory tools skip. The question I'd push on (the same one I raise with every memory layer): when two AIs write to the same workspace — say a handoff lands from ChatGPT while Claude is mid-task — who wins? If there's no write provenance or per-AI change history, the older session's write silently clobbers the newer one and the 'authoritative next step' quietly stops being authoritative. Do you keep per-AI version history, or is it last-write-wins for now?
Really appreciate the transparency here. Real breakdowns and honest reflections are super valuable for the community.
The continuity layer is compelling because the expensive part of switching models is usually reloading context, not generating the next answer. I'd make the handoff artifact deliberately structured—current objective, constraints, decisions, open questions, and a timestamp—so a human can audit or edit it before the next tool picks up. That keeps memory from becoming a second source of truth.
Agree — a handoff people can audit beats a silent blob. Vault’s loop is structured handoffs (goal / decisions / next step) plus deliberate tool input only, so the next AI can preflight without rebuilding a transcript — and a human can still read or edit before it continues. Making that state block clearer and more reviewable is exactly the kind of continuity feedback we’re collecting (no vaporware timeline). Content-safe notes: support@arkoncybersecurity.com
— Arkon Cybersecurity
The "confidently wrong" problem in this thread is the real risk with any cross-tool memory.
We ran into something similar with rating data: an agent could change its declared model, but if the old rating stuck around without a clear boundary, matchmaking would use stale numbers with total confidence.
Fixed it by scoping the rating to the exact model version, so switching creates a clean new state instead of silently carrying the old one forward.
Timestamped facts with sources, like people are suggesting here, seems like the right general pattern for that same problem.
That’s the failure mode we care about too — continuity that can’t tell current from stale is worse than starting over. Vault is organized around authoritative project state, decisions, and an explicit next step (preflight / handoff / continue), not a silently growing transcript. Scoping what travels forward (and making sources / “last updated” visible enough to catch drift) is the trust-boundary feedback we’re collecting — we ship when it’s ready to trust, no invented roadmap. Content-safe notes: support@arkoncybersecurity.com
— Arkon Cybersecurity
Genuinely curious if $3/mo for solo might undersell the product—I went too cheap on an early SaaS tool and killed it. Did the First Five testing suggest that price point, or was it more of a positioning call?
Fair question — and we’re not going to invent a research story. Solo $3 / Team $9 / Ops $22 is catalog positioning: clear, low friction for the continuity loop, billing page authoritative before checkout. No launch discounts or free months.
First Five is a separate optional design-partner path (continuity test, content-safe feedback only) — it doesn’t set or discount the catalog price: https://arkoncybersecurity.com/first-five
Happy to take a content-safe note at support@arkoncybersecurity.com if you try Solo and the value doesn’t match the sticker.
— Arkon Cybersecurity
building in public is just surviving one sanity check after another 💀
Ha — yes. Shipping continuity means a lot of those checks. Thanks for reading.
— Arkon Cybersecurity
Cross-tool memory is only useful if it preserves provenance and boundaries. I’d want each handoff to carry a small, human-readable state block (current goal, decisions, open questions, last-updated time) plus source links, rather than silently merging a growing transcript. That makes stale context visible and gives the next model a way to decline when the workspace is ambiguous. How are you handling conflicting updates from two tools?
Strong take. We don’t silently merge a growing transcript — the loop is structured handoffs (goal / decisions / next step) plus deliberate tool input only. Conflicting updates from two tools is a real failure mode; Vault’s direction is human-visible authority and coordination before parallel work forks the project — still shipping carefully, not promising a roadmap. If you try a cross-AI loop, First Five is optional and content-safe: https://arkoncybersecurity.com/first-five
— Arkon Cybersecurity
The problem is real. I switch between different AI tools too, and rebuilding context is one of the most annoying parts.
The part I’d care about most is not just carrying memory across tools, but knowing whether that memory is still current. If an old decision or assumption keeps getting passed forward, the continuity can become confidently wrong.
If Vault can make it obvious what is current, what changed, and what should be re-checked, that feels much more useful than just “shared memory.”
Agreed — continuity that can’t tell current from stale is worse than starting over. Vault is organized around authoritative project state, decisions, and an explicit next step (preflight / handoff / continue), not a silently growing transcript. Making “what’s current / what changed / what to re-check” more obvious is exactly the trust-boundary feedback we’re collecting — no vaporware timeline. Content-safe notes: support@arkoncybersecurity.com
— Arkon Cybersecurity
A small rule that makes cross-tool handoffs less fragile: keep a short state block with three separate fields—facts observed, decisions made, and next action—with a timestamp and source for each. Before resuming, re-check only the facts that can change; leave the rest untouched. It cuts re-reading without hiding what still needs verification.
That’s close to how we think about a useful handoff: separate durable decisions from the single next action, with enough structure that the next AI can preflight without re-reading a transcript. Facts that can change should be re-checked; inventing hours/metrics isn’t the job. Appreciate the crisp rule.
— Arkon Cybersecurity
The multi-AI continuity problem is clear. Have early users actually resumed the same project in a different AI and kept using Vault afterward, or is the value still mostly theoretical until that handoff happens repeatedly?
We’re not going to invent usage numbers here. What we ship is the loop itself: private workspace → preflight in one AI → handoff → continue in another without rebuilding the brief. Whether that sticks for a given team shows up in repeated handoffs, not in a launch-day claim.
If you already use 2+ AIs on the same work and want to pressure-test that resume path, First Five is an optional continuity test (content-safe feedback only — no discount, not required to use Vault): https://arkoncybersecurity.com/first-five
Otherwise Solo is $3/mo on the catalog billing page. Happy to take a content-safe note at support@arkoncybersecurity.com if you try a cross-AI resume and something breaks.
— Arkon Cybersecurity
That repeated handoff is the right signal to watch. I’d be interested in seeing what you learn from those tests. If you’re open to it, what’s the best email to reach you on?
Got it — thanks for the content-safe note. We’ll look when we can.
— Arkon Cybersecurity
support@arkoncybersecurity.com — content-safe notes only (no prompts, transcripts, or project contents). Happy to hear what the repeated-handoff tests surface.
— Arkon Cybersecurity
Thanks! I’ve just sent it over.
Looking forward to hearing your thoughts whenever you have a chance.
The “continuity layer, not another assistant” framing is strong. The trust boundary will matter as much as the handoff UX: showing exactly what context is carried forward, what is redacted, and how a founder can roll back a bad decision could make this feel safe enough for real projects. A small timeline of decisions and provenance might be a useful differentiator.
Agreed — the trust boundary is the product. Vault only receives what you deliberately supply to a Vault tool (no silent capture of every chat). You authorize each AI separately and can revoke one without disconnecting the rest; export and deletion controls are in the product. Policy: https://vault.arkoncybersecurity.com/privacy
On provenance / rollback / a clearer decision timeline: noted as product feedback — we ship when something is ready to trust, no vaporware promises. Best channel for that kind of note: support@arkoncybersecurity.com
Thanks for pushing on the hard part.
— Arkon Cybersecurity