*By using this super-intelligence tool you accept responsibility for the actions of your agents.
OPENCLAW + PROOF OF INTERACTION INCLUDED
Install once. Launch your agent.
Setup opens your local chat automatically. Choose your model and Timechain; PoI is already equipped. No separate bridge installation or pairing code.
Install ChainChat ↗
TIMECHAIN UNPAIRED
LEARNING —
Controls ↗
Stop current run ■
Your next interaction becomes part of your history.
YOUR ABILITY BAR OUTLINED = EQUIPPED · GLOWING = RUNNING
1 Recall EQUIPPED
2 Fork CONFIGURE
3 Continuum 1 TURN / BLOCK
4 Faculties LOADOUT
5 Verify CHAIN INTEGRITY
6 Covenant ACTIVE WITH CHAIN
7 Replay AUTOMATIC
8 Learning ENABLED
9 Dream MANUAL
0 Cognition MONITOR
02 / OPENCLAW CONTROL CENTER Your runtime. In reach. Configure native OpenClaw below your Cypher Tempre chat.
Back to chat ↑ Open your installed ChainChat app to load authenticated native controls.
Connect local app Local authentication, private credentials, required PoI enforcement and explicit tool grants remain operator-protected. Every native configuration family is discoverable in Settings; managed paths explain their restrictions.
Settings
Goals
Chat controls
Tool access
Channels
Schedules
Service activity
Plugins
Browser
Runtime
Authenticated native configuration
ChainChat frontend messages use followup: each request keeps its own identity and PoI review, with commitment following the selected Timechain block policy. Native /queue preferences apply to other native events and channel inputs; frontend requests stay separate.
Automatic native metadata utilities are off by default. Explicitly configuring agents.defaults.utilityModel requires a native-settings owner grant and uses that model’s own native provider credentials, billing and privacy terms, even while chat uses Free models. The shared free-tier key is not supplied. These metadata-only calls have no action tools; agent answers and actions retain canonical PoI review.
Configuration families Browse the complete native schema. Managed families stay visible with the reason they require their dedicated control or local operator.
Connect the local app to load its exact OpenClaw schema.
Native configuration Edit a JSON settings overlay against OpenClaw's own schema. Provider keys stay in the model connection panel. Required PoI and private-service settings remain managed.
Settings overlay Validate & apply settings Reload saved settings
Operator-protected settings and reasons
Native session goal A durable OpenClaw objective attached to this paired agent. Goal turns still pass through the Timechain's PoI gate and block policy.
Objective Start /goal Resume Pause Mark complete Clear goal
No selected native goal. Goals preserve their native state and usage limits across restarts. Starting or resuming admits a reviewed agent turn; it does not create an unbounded background loop.
OpenClaw chat controls Use native slash commands in the chat or here. Session preferences apply to the selected agent; resetting native chat context keeps its Timechain.
Conversation memory policy Connect the runtime to read its context policy. A model context window remains bounded.
The optional reviewed summary is an explicit advanced action. It passes through PoI and keeps the Timechain; it is separate from ordinary relevant-history retrieval.
Agent services Bind native channels and schedules to the selected verified Timechain. Each model turn and outgoing message passes through PoI. Work stays held while its owner session, connected model or chain is ineligible.
Retry eligible held work
Native runtime and session Gateway operator credentials are private to the local supervisor. The frontend never receives them.
CYPHER TEMPRE CHAINCHAT / LIVE BETA
Your agent. Your machine.
Launch ChainChat, choose your API model, and create a genesis or select an existing Timechain.
Your agent runs locally. Its model connection and Timechain stay under your control.
BLOCKSPACE / CHAIN AUDITOR
Inspect the chain beneath your agent.
Inspect your selected agent through the same local app. No second
connection or separate audit service is needed.
Audit selected Timechain
Open legacy auditor (advanced)
YOUR AGENT RUNS ON YOUR MACHINE
One install. Everything starts together.
The installer includes OpenClaw, Proof of Interaction, the Timechain engine and this chat interface. It starts the app and connects the local chat automatically.
Choose your system below. On macOS or Linux, extract the archive and open its installer.
In the chat it opens, choose your API model and enter its key locally.
Create a genesis or select an existing Timechain. Your equipped faculties and skill bar work there.
Already installed? Open ChainChat from your application launcher. It opens the authenticated local chat. Your API key stays in the running app's memory.
The website is also available as an optional remote-facing interface to your local app. That advanced mode uses a connection code; the installed chat does not.
Use website mode (advanced)
LOCAL OPENCLAW APP CONNECTED
Your local runtime is paired.
This browser can work with your local agents. End the session to
revoke its access and clear its model credential from app memory.
Your saved Timechains and pending blocks remain on this machine.
Show website pairing code
End app session
CHRONOSYNAPTIC FORKING
CONTINUUM · DATA HEIGHT
Set the rhythm of memory.
By default, each accepted input and output share one conversation block.
Turn on Continuum to accumulate or ingest larger blocks of data.
Manually ingest data +
THE GENESIS COVENANT
Alignment stays active
While Timechain is active, each generated response is held while a separate model review scores its
harmony with your agent’s genesis covenant. The Proof of Interaction
(PoI) gate must also accept it.
GENERATE → REVIEW → PoI GATE → SEAL & RELEASE
Low scores, missing scores, and failed reviews hold the response. Model
assessments can be imperfect; the gate enforces the supplied assessment
rather than guaranteeing factual accuracy or moral judgment.
Pausing real-time memory stops recall, learning and new chain recording.
Every response still requires PoI review. Action tools remain off until memory resumes.
CAPABILITIES TRAVEL
Share what your agent learns.
Exchange portable packs of modalities and senses. Each agent keeps its
own identity, history, and calibration. Preview each mechanism’s runtime
compatibility and the metadata it preserves before importing.
Receive a pack
Share faculties
Import reviewed faculties ↗
TIMECHAIN INSPECTOR
Your history, verifiable.
TIMECHAIN CONTROL & LIVE MONITOR
Your agent, at your pace.
Timechain active
Memory, faculties, covenant review and accepted conversation blocks are active.
Pause real-time memory · keep PoI enforced
LIVE ACTIVITY NOT CONNECTED
Pair a Timechain to see its actual runtime stages.
NATIVE MAINTENANCE
Enable learning and run these between conversations. They operate on the paired agent, use a model review, and may append native maintenance records.
Rebuild recall index Train learners Dream & consolidate
DELIBERATE DORMANCY
Pause real-time memory?
Memory, recall, faculty computations, learning and new chain blocks will pause. Every response still passes semantic and native PoI review. Action tools are off. Each request uses its current input without earlier chat context; existing history remains intact.
A current run will be stopped before pausing. An atomic block commit already underway may finish. Your selected provider still receives the input and candidate needed for generation and required review.
Pause memory · keep PoI enforced
FACULTY GROWTH & HIBERNATION
Choose what grows with you.
Inspect emerging faculties, promote supported proposals and choose what to equip. Pruning hibernates a faculty; its definition and past history remain available.
ShowAll learned faculties New this session Emerging Available Hibernating Refresh ↻