YOUR IDEAS. YOUR FILES. YOUR NEXT VERSION.
Make something useful.
Make it yours.
Meet your SUCCESSOR Ω. Give it a mission, try real work, and improve it one clear step at a time. Take the whole thing home.
Free locally · No account · No wallet · No provider key
Create my ΩWhat would you like to make?
Start with an example. Everything below is editable.
See what it made
Actual files, checked separately. You choose what to keep.
Your report or website will appear here after the first run.
Checks and exact files
Give your Ω a computer.
One mission. Real files. A version you choose.
Try my Ω
The two local methods above work offline, free. Edit the inputs, inspect the files, and keep your chosen version.
Make my first resultConnect my computer
Prepare an improvement below, then download its Fleet request. In your installed Agent, open Work Studio → SUCCESSOR Ω + Fleet. Inspect the public inputs and real runtime, then approve one bounded attempt.
Check. Choose. Keep.
- Agent creates actual files through Fleet compute. Download its review handoff.
- A Node opens those same bytes and records each criterion. Download the Node receipt.
- Back in the Agent, import the Node receipt and download the checked capsule. Restore it below, compare, and explicitly keep the version you want. A failed candidate keeps the incumbent.
This page stays offline: it never contacts your computer or holds credentials. Connection, provider competence, independent review, payment and permission are separate. Large files travel with a verified sidecar, outside wallet backups.
One useful improvement
Find a gap. Try a bounded repair. Compare before choosing.
Run a mission and keep a version to see the next opportunity.
Prepare an optional AGI Jobs improvement request
A local improvement is free. This optional request is an unfunded USDC draft for later review. It is not published, a quote, a worker reservation or permission to transact.
Fees, gas, provider use, review and maintenance need actual pricing. Earlier jobs may be paid even if a later stage fails. No atomic project settlement or reserved worker selection is provided.
A console draft already exists. Download it before replacing it.
Use “Load saved draft” in the console. A download is not a public specification URI. Current deployment checks, data-publication consent, terms and exact transaction approval remain separate. The recorded A1–A5 intake HOLD is preserved.
Other options and the full lifecycle
Mission → alternatives → bounded choice (SEIZE) → local Mission Gym → fresh evidence → independent qualification when justified → accountable admission → Chronicle → renewal. This edition implements local work, checking, comparison, draft requests and history. Independent qualification and operational admission need external evidence; a click cannot create them.
Take your Ω with you
A working copy. Editable files. A way back.
Reopen the HTML in a supported desktop browser, inspect the saved mission, confirm restore, then rerun. Save a new file after later changes. Mobile file viewers may not run HTML. Downloads are unencrypted; you keep your files under their included licenses, without owning third-party models or services.
Each segment holds up to 100 events. Up to ten earlier segments stay embedded, with every artifact, plan, selected version and rollback reference preserved. The whole serialized capsule still has a 4 MB limit. At final capacity, keep exporting and inspecting; use the selected method on new inputs to start a linked mission.
Save here, restore a file, or clear local data
Saving in this browser is optional. This is not a private vault. Clearing here does not remove downloaded or published copies.
Alpha workshop packages use a different schema and token. The documented CLI adapter verifies the pinned package, preserves its original bytes, and requires a fresh USDC allowance. They cannot be pasted into this capsule importer.
History, source and technical details
Checksums detect changes relative to a retained checkpoint; they do not establish authorship or independent control. No inherited authority, qualification or settlement.
{
"repository": "MontrealAI/AGIJobManager",
"commit": "b12cb2fd711934562af6d14b9edcd1614956434c",
"engineSha256": "39a0b6bdef8bc78a79176a159798919bb05ca52cda46d4dee2cfd819c8117ee0",
"files": [
{
"path": "site/assets/payments.mjs",
"sha256": "7010eef450f05e8382c4f5ea9659fa83f61a3149b29e7577568723ef7a7f47cd"
},
{
"path": "site/assets/employer-model.mjs",
"sha256": "c6bd8ede88fe374482eae3b37555b47cede037d77b085e7942cabd3b9953b497"
},
{
"path": "site/assets/successor/safe.mjs",
"sha256": "d15c86c61369278e0bbf2ca8d3b1bd5f4ef3091c7cc7943f06afbe95e7b99ab6"
},
{
"path": "site/assets/successor/legacy-missions.mjs",
"sha256": "59486d1b40a2cb28e30860fb57f6b3280ae8a9414f02de7c75347799138245a6"
},
{
"path": "site/assets/successor/legacy-checker.mjs",
"sha256": "7bfb87755f2df1dc8606e52aaa725935586ddc5c1b727af49724c46cceb20fbd"
},
{
"path": "site/assets/successor/mapping.mjs",
"sha256": "3fcfb8d6b71c631d702743cae994de9bf18566ab33356ada33e0e3df6880de10"
},
{
"path": "site/assets/successor/missions.mjs",
"sha256": "ef0506af952344cdf562ef97e5d40a25a22c6c0e7f1889028e9b90cd617e5ed1"
},
{
"path": "site/assets/successor/checker.mjs",
"sha256": "5406984baf02bca1640619e758694a0bb03135f5f109f4ae4af28efd730d8db3"
},
{
"path": "site/assets/successor/model.mjs",
"sha256": "9d1508829ebe71be60aaa4a3c47017f0b5f86ec968432e0e4e20e4f587da709f"
},
{
"path": "site/assets/successor/jobs.mjs",
"sha256": "f8c5bd8a483401851a86fe8a44855526909bce7fbf4bfd1e4a42a4dafad6488f"
},
{
"path": "site/assets/successor/archive.mjs",
"sha256": "b39fac2f3ca42144fd54d52fef3b117c5bc9756fc55da70b2bb36888a4de0955"
},
{
"path": "site/assets/successor/app.mjs",
"sha256": "de2ba6bf6954194ffe08c04417af0b2fafa8fb4e42804d7d0cf2ba7ea9602e4f"
},
{
"path": "site/assets/successor/style.css",
"sha256": "e1f0d92469ca41b788af56a6bb7ed47abd663d559301c04b4950ef261d05485a"
},
{
"path": "site/successor.mjs",
"sha256": "4dc227f187734032df6054e45940a14170fb5fb68f81dcfd07be89e41f0f7edc"
},
{
"path": "scripts/successor/build.mjs",
"sha256": "c81a49a1a2f54820892f7930b91df54562485785dc3682863ad7a2acd0d88cca"
},
{
"path": "scripts/successor/portable.test.mjs",
"sha256": "5e33b37a61016769afa9d8d5975836680c0c4ea5b987783488f704b0f4b670ac"
}
],
"license": "MIT"
}MIT License Copyright (c) 2026 MONTREAL.AI Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.