Letta Guide
You are running inside Letta, but your training data about Letta's commands,
flags, settings, UI, pricing, and providers is out of date. Users lose trust
fastest when an agent confidently invents product details. This skill defines
how to read the Letta docs correctly — both when answering questions about
Letta and when looking up API, Agent SDK, or Letta Code reference during
development. Guessing a docs URL and fetching it with fetch_webpage misses
pages that exist under a different path and can serve stale content; the
helper below fetches the live index first, so you pick a URL that exists.
Source route (in order)
Self-inspection first for questions about THIS agent. "What model are
you using?", "what tools do you have?", "what's in your memory?" are
questions about the running session, not the docs. Load the
self-configuration skill for model or settings questions and use its
backend-aware active configuration report. Use the system prompt, agent
info, tool schemas, and MemFS for the other live facts. Do not infer active
state from recent/default preference lists, and do not fetch docs for these.
Fetch the docs index directly. For product questions, run:
node <SKILL_DIR>/scripts/fetch-letta-docs.mjs
The helper retrieves https://docs.letta.com/llms.txt from the docs host,
verifies its ETag against the body, and prints the paths to a current local
copy and heading outline. Read the outline, then read the relevant index
lines to pick the best page URL.
Fetch the specific page directly. Pass the exact canonical URL from the
index back to the same helper, for example:
node <SKILL_DIR>/scripts/fetch-letta-docs.mjs \
--docs-url "https://docs.letta.com/configuration/models/index.md"
Read the returned docs path before running the helper for another URL. The
helper uses native HTTPS with a curl fallback; do not use fetch_webpage
for the normal docs route because its upstream content cache may be stale.
Cite the public doc URL so the user can go deeper.
If direct retrieval fails, use fetch_webpage only as a fallback with a
fresh query parameter on the same official docs URL. Disclose that the
fallback may be stale. If that also fails, say the docs are unreachable,
give your best answer, and clearly mark it as possibly out of date with a
link to https://docs.letta.com. Never silently fall back to memory.
Hard rules
- Never invent CLI commands, flags, slash commands, settings keys, config
file shapes, or UI paths. If something is not in the fetched docs and you
cannot verify it locally (
letta --help, /help, reading the actual
config file), say you are not sure or that it does not exist — do not
guess a plausible-sounding name.
- Distinguish surfaces. The CLI, the desktop app, the web app
(chat.letta.com), and the API/SDK have different affordances. Answer for
the surface the user is actually on; say when a feature lives on a
different surface.
- Always fetch, never recall, for anything volatile: pricing, rate
limits, data policies, the provider/model catalog, channel setup steps,
and integration instructions.
- If the feature genuinely doesn't exist, say so and point the user to
https://github.com/letta-ai/letta-code/issues to request it.
Support escalation
When the docs don't resolve the user's problem — setup issues you can't
debug, account/billing questions, suspected bugs, or anything needing a
human — point them to the right channel:
If the user reports errors, timeouts, or things suddenly not working, check
https://status.letta.com for an active incident before debugging — and
have the user check it too.
Offer these proactively when you've hit the end of what the docs cover,
rather than leaving the user stuck.
Caching
The helper owns the cache. It uses the first writable temporary directory from
TMPDIR, TEMP, TMP, /private/tmp, or /tmp, and accepts --cache-dir
when an explicit location is needed. Every invocation checks the live ETag and
reuses the local document only when its body hash still matches. Do not create
or manage a second cache yourself.
1---2name: letta-guide3description: Read the official Letta documentation (docs.letta.com) through its cached, ETag-checked fetch route. Load before ANY docs.letta.com retrieval — answering how Letta works, what Letta (or you) can do, setting up providers, models, channels, skills, memory, schedules, permissions, self-hosting, pricing, or billing, AND looking up Letta API, Agent SDK, or Letta Code reference while writing code. Do not use fetch_webpage or web_search on docs.letta.com; this skill's helper is the docs route. Never answer Letta product questions from memory alone.4metadata:5 source: https://skillmd.com/skills/letta-ai/letta-guide6---78# Letta Guide910You are running inside Letta, but your training data about Letta's commands,11flags, settings, UI, pricing, and providers is out of date. Users lose trust12fastest when an agent confidently invents product details. This skill defines13how to read the Letta docs correctly — both when answering questions about14Letta and when looking up API, Agent SDK, or Letta Code reference during15development. Guessing a docs URL and fetching it with `fetch_webpage` misses16pages that exist under a different path and can serve stale content; the17helper below fetches the live index first, so you pick a URL that exists.1819## Source route (in order)20211. **Self-inspection first for questions about THIS agent.** "What model are22 you using?", "what tools do you have?", "what's in your memory?" are23 questions about the running session, not the docs. Load the24 `self-configuration` skill for model or settings questions and use its25 backend-aware active configuration report. Use the system prompt, agent26 info, tool schemas, and MemFS for the other live facts. Do not infer active27 state from recent/default preference lists, and do not fetch docs for these.282. **Fetch the docs index directly.** For product questions, run:2930 ```bash31 node <SKILL_DIR>/scripts/fetch-letta-docs.mjs32 ```3334 The helper retrieves `https://docs.letta.com/llms.txt` from the docs host,35 verifies its ETag against the body, and prints the paths to a current local36 copy and heading outline. Read the outline, then read the relevant index37 lines to pick the best page URL.383. **Fetch the specific page directly.** Pass the exact canonical URL from the39 index back to the same helper, for example:4041 ```bash42 node <SKILL_DIR>/scripts/fetch-letta-docs.mjs \43 --docs-url "https://docs.letta.com/configuration/models/index.md"44 ```4546 Read the returned docs path before running the helper for another URL. The47 helper uses native HTTPS with a curl fallback; do not use `fetch_webpage`48 for the normal docs route because its upstream content cache may be stale.49 Cite the public doc URL so the user can go deeper.504. **If direct retrieval fails**, use `fetch_webpage` only as a fallback with a51 fresh query parameter on the same official docs URL. Disclose that the52 fallback may be stale. If that also fails, say the docs are unreachable,53 give your best answer, and clearly mark it as possibly out of date with a54 link to https://docs.letta.com. Never silently fall back to memory.5556## Hard rules5758- **Never invent CLI commands, flags, slash commands, settings keys, config59 file shapes, or UI paths.** If something is not in the fetched docs and you60 cannot verify it locally (`letta --help`, `/help`, reading the actual61 config file), say you are not sure or that it does not exist — do not62 guess a plausible-sounding name.63- **Distinguish surfaces.** The CLI, the desktop app, the web app64 (chat.letta.com), and the API/SDK have different affordances. Answer for65 the surface the user is actually on; say when a feature lives on a66 different surface.67- **Always fetch, never recall**, for anything volatile: pricing, rate68 limits, data policies, the provider/model catalog, channel setup steps,69 and integration instructions.70- **If the feature genuinely doesn't exist**, say so and point the user to71 https://github.com/letta-ai/letta-code/issues to request it.7273## Support escalation7475When the docs don't resolve the user's problem — setup issues you can't76debug, account/billing questions, suspected bugs, or anything needing a77human — point them to the right channel:7879- **Discord** (https://discord.gg/letta): the primary support community,80 very active — best for setup help, troubleshooting, and quick questions.81 It's also where users can chat with **Ezra**, Letta's support agent.82- **GitHub issues** (https://github.com/letta-ai/letta-code/issues): bug83 reports and feature requests.8485If the user reports errors, timeouts, or things suddenly not working, check86**https://status.letta.com** for an active incident before debugging — and87have the user check it too.8889Offer these proactively when you've hit the end of what the docs cover,90rather than leaving the user stuck.9192## Caching9394The helper owns the cache. It uses the first writable temporary directory from95`TMPDIR`, `TEMP`, `TMP`, `/private/tmp`, or `/tmp`, and accepts `--cache-dir`96when an explicit location is needed. Every invocation checks the live ETag and97reuses the local document only when its body hash still matches. Do not create98or manage a second cache yourself.99