What is Ancient Echoes AI Bazi?

Ancient Echoes combines advanced AI with traditional Bazi charting to provide personality analysis and chart-based reflection insights.

What classics can I read here?

We offer a vast collection of Taoist, Buddhist, TCM classics, and modern spiritual works.

Is Ancient Echoes like an Eastern alternative to MBTI?

In search-intent terms, yes, it can feel like an Eastern alternative to MBTI. But instead of a pure personality label system, BaZi uses birth-chart structure, five-element balance, relationship patterns, and life timing to guide interpretation.

How is BaZi different from MBTI or horoscope apps?

MBTI focuses on personality types, while horoscope apps often give broad daily content. BaZi works from your chart structure, useful elements, and life-cycle timing, making it closer to a chart-based personality and life-timing reading system.

What other systems does Ancient Echoes support besides BaZi?

Ancient Echoes now also supports Zi Wei Dou Shu, Seven Luminaries, Western astrology, Tarot, and Liuyao divination, each with AI follow-up grounded in the active context.

How is the AI here different from generic spiritual chat?

The main difference is context grounding: the AI is designed to keep answering from the current chart, spread, or divination state instead of drifting into generic spiritual talk.

LingXi private assistant

Bring Ancient Echoes Q&A to Telegram, Feishu, OpenClaw, and server terminals with the same account, RAG, quota, and audit controls.

Open my LingXi devices View plans and quota

Why a controlled gateway

This is not an open public relay. It reuses on-site identity, logs, and the RAG control plane.

Device auth reuses the website account system: JWT / Cookie / sk- / lxd_.

RAG still runs through the website backend on private hosts and is not exposed to terminals.

Compatible with chat/completions, responses, audit logs, and device config delivery.

1. One-command install

If OpenClaw is missing, the installer uses the official npm package, then prepares local directories, a persistent installer script, and status files. It does not rewrite a live OpenClaw config immediately.

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash

2. Cloud servers: API key init

Headless, CI, and long-running nodes should prefer sk- keys. Device center can also mint a long-lived on-site key.

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash -s -- init --base-url https://www.ancientechoes.ai --auth-token "$LINGXI_AUTH_TOKEN"

3. Personal devices: device flow

The terminal prints a user_code and verification_uri_complete. Approve it in a browser on another device, or run install first and reuse the local script to continue linking.

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash -s -- link --base-url https://www.ancientechoes.ai

Two recommended paths

Do not force every setup through a browser approval flow. Servers and personal devices should be handled separately.

Headless / cloud server

Prefer an on-site sk- API key and initialize config directly. Long-running nodes can create a permanent key in device center, with no desktop or local browser required.

Personal terminal / short-lived node

Use device flow. The terminal prints a user_code; approve it in a signed-in browser elsewhere, then the device exchanges an lxd_ token.

Plans and access notes

These are LingXi usage notes, not a replacement for live prices on the upgrade page.

Free

Trial

Finish device auth and a single-entry dry run

For website users who are just starting with LingXi

Get install / link / doctor working first

Good for verifying OpenClaw or a single channel

Live quota and price always follow the upgrade page

Pro

Recommended

Default plan for stable Q&A and knowledge-base use

For people who keep Telegram / Feishu / OpenClaw running

Better for ongoing chat/completions and responses

Covers regular RAG Q&A and device management

The suggested default for daily LingXi use

View Pro plan

Master

High usage

For high-frequency Q&A, many devices, and heavy coding

For coding-assistant and high-activity private-assistant workflows

Better for frequent coding and always-on entry points

For users sensitive to model routing, quota, and response rate

Intended for intensive private-assistant work

View Master plan

Channel status

The first release focuses on Telegram / Feishu / OpenClaw. DingTalk stays on a bridge path.

Supported

The installer can write config and call native OpenClaw commands.

Supported

Feishu / Lark

Merges config files, then validate + gateway start, without an interactive wizard.

Supported

OpenClaw Web

Can start a local gateway on its own and use the website API for unified RAG and auth.

Bridge preview

DingTalk

Currently an external bridge. Not a first-release promise; live webhook wiring still needs more work.

Device-side API examples

The repo already has a runnable CLI and reusable SDK for device or terminal integration.

LingXi CLI: chat

Best for conversational Q&A, RAG retrieval, and persona-based daily assistants.

LINGXI_AUTH_TOKEN=sk-xxx node scripts/lingxi-cli.js chat --prompt "请总结今天的运势重点" --stream

The CLI reuses the separately packable services/lingxi-sdk-js package. To distribute it outside this repo, run npm pack in that directory.

View device status View plan quota

LingXi CLI: responses

Best for coding, instruction-style generation, and OpenAI Responses-like access.

LINGXI_AUTH_TOKEN=sk-xxx node scripts/lingxi-cli.js responses --scene coding --prompt "写一个 Go 函数，判断字符串是否是回文" --stream

If you do not want the CLI, install or import @ancientechoes/lingxi-sdk in your own script, or follow scripts/lingxi-api-example.js.

View device status View plan quota

FAQ

Current product boundary, auth model, and first-release scope.

Is RAG local?

No. The first release still forwards through the Ancient Echoes website backend to a private/cloud RAG service. Terminals never receive the raw RAG address.

Do I need a separate LingXi account?

No. LingXi reuses the on-site account system. The device side only turns website auth into a browser-free device export.

Can a cloud server without a desktop connect?

Yes. Prefer an sk- API key with init --auth-token. Device center can mint a permanent key. If you would rather bind a device than issue a long-lived token, run link on the server and approve it from another machine with a browser.

Do I have to install OpenClaw?

No. OpenClaw is the most complete first-release local runtime, but canonical /api/v1/ai/* can also be called from your own scripts, CLI, or other clients.

Can it be used for coding?

Yes. Canonical /api/v1/ai/* already supports chat and coding scenes, and responses has a minimal compatible path.

Is this an open public relay?

No. It remains a controlled compatibility gateway with on-site auth, quota, logs, device grants, and the RAG control plane.

What is the DingTalk status?

It currently uses an external bridge. The protocol baseline exists, but it is not a first-release promise. More event types and live credentials still need work.

What to do next

If you are ready to connect, this is the shortest path.

Run install + link with the installer, then approve the device on the website.

Then pick start-telegram, start-feishu, or start-openclaw for your channel.

For device-side integration, run scripts/lingxi-cli.js or reuse services/lingxi-sdk-js in your own script.

To inspect devices or revoke an old node, return to My LingXi devices.

Go to device center

LingXi private assistant

Bring Ancient Echoes Q&A to Telegram, Feishu, OpenClaw, and server terminals with the same account, RAG, quota, and audit controls.

Open my LingXi devices View plans and quota

Why a controlled gateway

This is not an open public relay. It reuses on-site identity, logs, and the RAG control plane.

Device auth reuses the website account system: JWT / Cookie / sk- / lxd_.

RAG still runs through the website backend on private hosts and is not exposed to terminals.

Compatible with chat/completions, responses, audit logs, and device config delivery.

1. One-command install

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash

2. Cloud servers: API key init

Headless, CI, and long-running nodes should prefer sk- keys. Device center can also mint a long-lived on-site key.

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash -s -- init --base-url https://www.ancientechoes.ai --auth-token "$LINGXI_AUTH_TOKEN"

3. Personal devices: device flow

The terminal prints a user_code and verification_uri_complete. Approve it in a browser on another device, or run install first and reuse the local script to continue linking.

curl -fsSL https://www.ancientechoes.ai/install-lingxi.sh | bash -s -- link --base-url https://www.ancientechoes.ai

Two recommended paths

Do not force every setup through a browser approval flow. Servers and personal devices should be handled separately.

Headless / cloud server

Prefer an on-site sk- API key and initialize config directly. Long-running nodes can create a permanent key in device center, with no desktop or local browser required.

Personal terminal / short-lived node

Use device flow. The terminal prints a user_code; approve it in a signed-in browser elsewhere, then the device exchanges an lxd_ token.

Plans and access notes

These are LingXi usage notes, not a replacement for live prices on the upgrade page.

Free

Trial

Finish device auth and a single-entry dry run

For website users who are just starting with LingXi

Get install / link / doctor working first

Good for verifying OpenClaw or a single channel

Live quota and price always follow the upgrade page

Pro

Recommended

Default plan for stable Q&A and knowledge-base use

For people who keep Telegram / Feishu / OpenClaw running

Better for ongoing chat/completions and responses

Covers regular RAG Q&A and device management

The suggested default for daily LingXi use

View Pro plan

Master

High usage

For high-frequency Q&A, many devices, and heavy coding

For coding-assistant and high-activity private-assistant workflows

Better for frequent coding and always-on entry points

For users sensitive to model routing, quota, and response rate

Intended for intensive private-assistant work

View Master plan

Channel status

The first release focuses on Telegram / Feishu / OpenClaw. DingTalk stays on a bridge path.

Supported

The installer can write config and call native OpenClaw commands.

Supported

Feishu / Lark

Merges config files, then validate + gateway start, without an interactive wizard.

Supported

OpenClaw Web

Can start a local gateway on its own and use the website API for unified RAG and auth.

Bridge preview

DingTalk

Currently an external bridge. Not a first-release promise; live webhook wiring still needs more work.

Device-side API examples

The repo already has a runnable CLI and reusable SDK for device or terminal integration.

LingXi CLI: chat

Best for conversational Q&A, RAG retrieval, and persona-based daily assistants.

LINGXI_AUTH_TOKEN=sk-xxx node scripts/lingxi-cli.js chat --prompt "请总结今天的运势重点" --stream

The CLI reuses the separately packable services/lingxi-sdk-js package. To distribute it outside this repo, run npm pack in that directory.

View device status View plan quota

LingXi CLI: responses

Best for coding, instruction-style generation, and OpenAI Responses-like access.

LINGXI_AUTH_TOKEN=sk-xxx node scripts/lingxi-cli.js responses --scene coding --prompt "写一个 Go 函数，判断字符串是否是回文" --stream

If you do not want the CLI, install or import @ancientechoes/lingxi-sdk in your own script, or follow scripts/lingxi-api-example.js.

View device status View plan quota

FAQ

Current product boundary, auth model, and first-release scope.

Is RAG local?

No. The first release still forwards through the Ancient Echoes website backend to a private/cloud RAG service. Terminals never receive the raw RAG address.

Do I need a separate LingXi account?

No. LingXi reuses the on-site account system. The device side only turns website auth into a browser-free device export.

Can a cloud server without a desktop connect?

Do I have to install OpenClaw?

No. OpenClaw is the most complete first-release local runtime, but canonical /api/v1/ai/* can also be called from your own scripts, CLI, or other clients.

Can it be used for coding?

Yes. Canonical /api/v1/ai/* already supports chat and coding scenes, and responses has a minimal compatible path.

Is this an open public relay?

No. It remains a controlled compatibility gateway with on-site auth, quota, logs, device grants, and the RAG control plane.

What is the DingTalk status?

It currently uses an external bridge. The protocol baseline exists, but it is not a first-release promise. More event types and live credentials still need work.

What to do next

If you are ready to connect, this is the shortest path.

Run install + link with the installer, then approve the device on the website.

Then pick start-telegram, start-feishu, or start-openclaw for your channel.

For device-side integration, run scripts/lingxi-cli.js or reuse services/lingxi-sdk-js in your own script.

To inspect devices or revoke an old node, return to My LingXi devices.

Go to device center