Developer Docs

Build on the Developers Digest platform

One API key, one credit balance. Generate images and voice, count tokens, resolve favicons, query model pricing, and pull the Skill, Agent, and Design libraries - over REST, MCP, or the dd CLI.

Get started Browse the API OpenAPI spec

Overview

What you can build

The platform exposes three surfaces over the same credit-metered capabilities. Pick whichever fits how you work.

REST API

https://www.developersdigest.tech/api/v1

Plain HTTP + JSON. Bearer auth. Every endpoint has a curl example below.

MCP endpoint

https://www.developersdigest.tech/api/mcp

Streamable HTTP MCP server. Drop it into any agent that speaks MCP.

dd CLI

dd generate image ...

Single-binary CLI for generation, skills, agents, and balance. Download at /dashboard/cli.

Prefer a visual workspace? The same credit-metered capabilities power Chat, Image generation, and Voice generation.

Authentication

API keys & credits

-Create an API key in the dashboard at /dashboard/keys. Keys are prefixed dd_live_.
-Send it on every request as an Authorization: Bearer dd_live_... header.
-A signed-in browser session also authenticates the /api/v1 REST routes; the MCP endpoint requires a real dd_live_ key (there is no session on that transport).
-Credits are shared with your account. Any spend from the API, MCP, or CLI draws from the same universal balance, and owner accounts are never charged.
-Metered actions only deduct credits on success. A provider failure returns an error and leaves your balance untouched.

curl

curl https://www.developersdigest.tech/api/v1/me \
  -H "Authorization: Bearer dd_live_your_key_here"

Create and manage keys at /dashboard/keys. Add credits at /pricing.

REST API

/api/v1 endpoints

All endpoints require a valid key or session and return plain JSON. Metered actions are charged only on success.

POST/api/v1/images1-6 credits

Generate an image from a text prompt. Models route through the Vercel AI Gateway; each model carries its own credit cost, resolved server-side.

JSON body

Field	Type	Req	Description
prompt	string	yes	Image description, 3 to 2000 characters.
size	string	no	One of square, square_hd, portrait, landscape, wide. Defaults to square_hd.
model	string	no	Image model id, e.g. xai/grok-imagine-image (default, 1 credit), openai/gpt-image-2 (3), google/gemini-3-pro-image (6). Invalid ids return 400 with the full list.
provider	string	no	Optional provider override: "gateway" (default) or "fal" (legacy, only when configured).

Response

Field	Type	Req	Description
url	string	yes	Durable image URL (falls back to the ephemeral provider URL when persistence fails).
persisted	boolean	yes	True when the image was saved to your gallery.
persistError	string	no	Present when the gallery save failed; the url is then ephemeral.
model	string	yes	Model that produced the image.
creditsSpent	number	yes	Credits charged (0 for owner accounts).
balance	number	yes	Remaining credit balance.

curl

curl -X POST https://www.developersdigest.tech/api/v1/images \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a hard-edged neutral workflow board","size":"landscape"}'

Response

{
  "url": "https://blob.../image.png",
  "persisted": true,
  "model": "xai/grok-imagine-image",
  "creditsSpent": 1,
  "balance": 999
}

-Rate limited to 10 requests / 5 minutes per account.
-Credits are deducted only after a successful generation, at the chosen model's rate.
-Returns 503 not_configured if image generation is not enabled on the server.

POST/api/v1/voice1 credit

Synthesize speech from text via the Vercel AI Gateway.

JSON body

Field	Type	Req	Description
text	string	yes	Text to speak, 3 to 2000 characters.
voice	string	no	One of alloy, echo, fable, onyx, nova, shimmer. Defaults to alloy.
model	string	no	One of openai/tts-1, openai/tts-1-hd. Defaults to openai/tts-1.

Response

Field	Type	Req	Description
audioBase64	string	yes	Base64-encoded MP3 audio (mime audio/mpeg).
url	string \| null	yes	Durable gallery URL for the clip, or null when persistence failed.
persisted	boolean	yes	True when the clip was saved to your gallery.
persistError	string	no	Present when the gallery save failed; play back via the audioBase64 data instead.
model	string	yes	TTS model used.
voice	string	yes	Voice used.
creditsSpent	number	yes	Credits charged (0 for owner accounts).
balance	number	yes	Remaining credit balance.

curl

curl -X POST https://www.developersdigest.tech/api/v1/voice \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"text":"Welcome to Developers Digest.","voice":"nova"}'

Response

{
  "audioBase64": "SUQzBAAAAAA...",
  "url": "https://blob.../voice.mp3",
  "persisted": true,
  "model": "openai/tts-1",
  "voice": "nova",
  "creditsSpent": 1,
  "balance": 999
}

-Rate limited to 10 requests / 5 minutes per account.
-Audio is returned as base64 JSON (not a raw audio body) so credits and balance travel with it. Wrap it in a data: URL for playback or download.
-Credits are deducted only after a successful generation.

POST/api/v1/text1-3 credits

Non-streaming text generation via the Vercel AI Gateway. Built for the dd CLI and scripts that want a single JSON completion back.

JSON body

Field	Type	Req	Description
prompt	string	yes	The prompt, 1 to 24,000 characters.
model	string	no	Whitelisted chat model id, e.g. anthropic/claude-sonnet-4.6 (2 credits) or openai/gpt-5.6-sol (3). Unknown ids fall back to the default (google/gemini-3.1-flash-lite, 1 credit).
system	string	no	Optional system prompt override, up to 8,000 characters.

Response

Field	Type	Req	Description
text	string	yes	The generated completion.
model	string	yes	Model that produced the text.
creditsSpent	number	yes	Credits charged (0 for owner accounts).
balance	number	yes	Remaining credit balance.

curl

curl -X POST https://www.developersdigest.tech/api/v1/text \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Summarize the MCP spec in three bullets."}'

Response

{
  "text": "- MCP is an open protocol...",
  "model": "google/gemini-3.1-flash-lite",
  "creditsSpent": 1,
  "balance": 999
}

-Rate limited to 20 requests / 5 minutes per account.
-Priced by the chosen model's credit tier; credits are deducted only after a successful generation.
-Returns 503 not_configured if text generation is not enabled on the server.

POST/api/v1/video3-18 credits

Generate a short video from a text prompt (and an optional input image for image-to-video models). Each model carries its own credit cost.

JSON body

Field	Type	Req	Description
prompt	string	yes	Video description, 3 to 2000 characters.
model	string	no	Video model id, e.g. bytedance/seedance-2.0-fast (default, 8 credits) or klingai/kling-v2.6-i2v (5). Invalid ids return 400 with the full list.
imageUrl	string	no	http(s) image URL. Required for image-to-video (i2v) models; used as the first frame by t2v models that support it.

Response

Field	Type	Req	Description
url	string	yes	Durable video URL.
model	string	yes	Model that produced the video.
creditsSpent	number	yes	Credits charged (0 for owner accounts).
balance	number	yes	Remaining credit balance.

curl

curl -X POST https://www.developersdigest.tech/api/v1/video \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a slow pan across a neutral workflow board"}'

Response

{
  "url": "https://blob.../video.mp4",
  "model": "bytedance/seedance-2.0-fast",
  "creditsSpent": 8,
  "balance": 992
}

-Rate limited to 5 requests / 5 minutes per account.
-Generation is slow; the route polls the provider and can take up to 3 minutes. Set your client timeout accordingly.
-Credits are deducted only after a successful generation, at the chosen model's rate.

GET/api/v1/faviconsFree

Resolve a company favicon across public providers (Google s2, DuckDuckGo, /favicon.ico) and report which URLs serve an image.

Query string

Field	Type	Req	Description
domain	string	no	Explicit domain, e.g. example.com. Validated for shape.
q	string	no	Company name to guess a domain from (best-effort, flagged in the response). Provide domain or q.

Response

Field	Type	Req	Description
domain	string	yes	Resolved domain.
guessed	boolean	yes	True when derived from q.
sources	array	yes	Each candidate { provider, url, size?, ok } with ok=true when it serves an image.
best	string \| null	yes	First provider URL that resolved, or null.

curl

curl "https://www.developersdigest.tech/api/v1/favicons?domain=vercel.com" \
  -H "Authorization: Bearer dd_live_your_key_here"

Response

{
  "domain": "vercel.com",
  "guessed": false,
  "sources": [
    { "provider": "google", "url": "https://www.google.com/s2/favicons?domain=vercel.com&sz=256", "size": 256, "ok": true }
  ],
  "best": "https://www.google.com/s2/favicons?domain=vercel.com&sz=64"
}

-Rate limited to 30 requests / 5 minutes per account.
-Only public provider URLs are checked and returned; image bytes are never proxied.

POST/api/v1/tokens/countFree

Count tokens in a string using OpenAI BPE encodings (o200k_base by default, cl100k_base for GPT-4 / GPT-3.5).

JSON body

Field	Type	Req	Description
text	string	yes	Text to tokenize (max 2,000,000 characters).
model	string	no	Model id to pick the encoding. Defaults to gpt-4o.

Response

Field	Type	Req	Description
tokens	number	yes	Token count.
model	string	yes	Model used to pick the encoding.
encoding	string	yes	o200k_base or cl100k_base.

curl

curl -X POST https://www.developersdigest.tech/api/v1/tokens/count \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"text":"Count these tokens.","model":"gpt-4o"}'

Response

{
  "tokens": 4,
  "model": "gpt-4o",
  "encoding": "o200k_base"
}

GET/api/v1/pricingFree

Query the kept-current AI model pricing and capability dataset. Built for agents that reason about cost and context before dispatching a call.

Query string

Field	Type	Req	Description
provider	string	no	Filter by provider, e.g. anthropic.
family	string	no	Substring match on family, e.g. sonnet.
id	string	no	Exact model id, e.g. claude-sonnet-5.

Response

Field	Type	Req	Description
models	array	yes	Matching ModelPricing entries.
updatedAt	string	yes	When the dataset was last updated.
count	number	yes	Number of matching models.

curl

curl "https://www.developersdigest.tech/api/v1/pricing?provider=anthropic&family=sonnet" \
  -H "Authorization: Bearer dd_live_your_key_here"

Response

{
  "models": [ { "id": "claude-sonnet-5", "provider": "anthropic", "family": "sonnet", "...": "..." } ],
  "updatedAt": "2026-06-01",
  "count": 1
}

-Rate limited to 60 requests / 5 minutes per account. Filters are AND-combined and case-insensitive.

GET/api/v1/meFree

Identity and balance for the authenticated caller.

Response

Field	Type	Req	Description
clerkId	string	yes	The caller's account id.
balance	number	yes	Current credit balance.
isOwner	boolean	yes	True when the account is never charged.
plan	string	yes	Current plan.
planRenewsAt	string \| null	yes	ISO renewal date, or null.

curl

curl https://www.developersdigest.tech/api/v1/me \
  -H "Authorization: Bearer dd_live_your_key_here"

Response

{
  "clerkId": "user_xxx",
  "balance": 1000,
  "isOwner": false,
  "plan": "pro",
  "planRenewsAt": "2026-08-01T00:00:00.000Z"
}

GET/api/v1/memoryFree

List or search your saved memories (notes and links).

Query string

Field	Type	Req	Description
q	string	no	Search term. Omit to list all memories, newest first.

Response

Field	Type	Req	Description
memories	array	yes	Each { id, kind, title, body, url, meta, createdAt }.

curl

curl "https://www.developersdigest.tech/api/v1/memory?q=mcp" \
  -H "Authorization: Bearer dd_live_your_key_here"

Response

{
  "memories": [
    { "id": "mem_1", "kind": "note", "title": "MCP config", "body": "...", "url": null, "createdAt": "..." }
  ]
}

POST/api/v1/memoryFree

Save a note or a link to your memory canvas.

JSON body

Field	Type	Req	Description
kind	string	no	Either "note" (default) or "link".
title	string	no	Optional title.
body	string	no	Note text. A note needs a title or body.
url	string	no	Required for a link; must be an http(s) URL.

Response

Field	Type	Req	Description
memory	object	yes	The created memory row. Returned with HTTP 201.

curl

curl -X POST https://www.developersdigest.tech/api/v1/memory \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"kind":"note","title":"MCP config","body":"basePath is /api"}'

Response

{
  "memory": { "id": "mem_2", "kind": "note", "title": "MCP config", "body": "basePath is /api" }
}

GET/api/v1/galleryFree

List your dashboard gallery assets, newest first.

Query string

Field	Type	Req	Description
limit	number	no	Max assets to return.
before	string	no	ISO date cursor for pagination.

Response

Field	Type	Req	Description
assets	array	yes	Each { id, kind, url, name, prompt, model, contentType, creditsSpent, createdAt }.

curl

curl "https://www.developersdigest.tech/api/v1/gallery?limit=10" \
  -H "Authorization: Bearer dd_live_your_key_here"

Response

{
  "assets": [
    { "id": "a_1", "kind": "image", "url": "https://.../image.png", "creditsSpent": 5, "createdAt": "..." }
  ]
}

POST/api/v1/galleryFree

Persist a just-generated asset into durable Blob storage. The generating call already charged, so persisting is free.

JSON body

Field	Type	Req	Description
kind	string	yes	Either "image" or "voice".
sourceUrl	string	no	For image: a fal.media / fal.run / fal.ai URL or a data:image URL, re-hosted server-side (SSRF-guarded).
audioBase64	string	no	For voice: base64 MP3 bytes.
prompt	string	no	Prompt to record for display.
model	string	no	Model to record for display.
creditsSpent	number	no	Credits the source generation cost, echoed for display.

Response

Field	Type	Req	Description
asset	object	yes	The stored asset row. Returned with HTTP 201.

curl

curl -X POST https://www.developersdigest.tech/api/v1/gallery \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"kind":"image","sourceUrl":"https://fal.media/files/example.png","creditsSpent":5}'

Response

{
  "asset": { "id": "a_2", "kind": "image", "url": "https://blob.../image.png", "creditsSpent": 5 }
}

-Image sourceUrl hosts are whitelisted to Fal (fal.media, *.fal.media, *.fal.run, *.fal.ai) or a data:image URL to prevent SSRF.
-Returns 503 not_configured when media storage is not enabled.

POST/api/v1/uploadsFree

Upload a member file (image or audio) into the dashboard gallery.

Multipart form

Field	Type	Req	Description
file	file	yes	A single image/* or audio/* file. Max 15 MB.

Response

Field	Type	Req	Description
asset	object	yes	The stored asset row. Returned with HTTP 201.

curl

curl -X POST https://www.developersdigest.tech/api/v1/uploads \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -F "file=@./cover.png"

Response

{
  "asset": { "id": "a_3", "kind": "upload", "url": "https://blob.../cover.png", "creditsSpent": 0 }
}

-Rate limited to 20 uploads / 5 minutes per account.
-Only image/* and audio/* are accepted; the file field is required.
-Returns 503 not_configured when media storage is not enabled.

GET/api/v1/skillsFree

The lean index of the Skill Library: slug, one-line description, and category for every skill. The entry point for progressive disclosure.

Response

Field	Type	Req	Description
skills	array	yes	Each { slug, description, category }.

curl

curl https://www.developersdigest.tech/api/v1/skills

Response

{
  "skills": [
    { "slug": "clerk-nextjs", "description": "Wire Clerk auth into a Next.js app.", "category": "auth" }
  ]
}

-Public and unauthenticated: the Skill Library is open content, cacheable for an hour.
-Scan this cheap list, then GET /api/v1/skills/{slug} for the overview and file manifest.

GET/api/v1/skills/{slug}Free

A skill's overview manifest: identity plus the file tree (paths and purposes, not contents) and a download link.

Response

Field	Type	Req	Description
slug	string	yes	Skill slug.
name	string	yes	Display name.
description	string	yes	One-line description.
category	string	yes	Skill category.
files	array	yes	Each { path, purpose }. Contents come from the download zip.
downloadUrl	string	yes	Absolute URL of the zip download.

curl

curl https://www.developersdigest.tech/api/v1/skills/clerk-nextjs

Response

{
  "slug": "clerk-nextjs",
  "name": "Clerk + Next.js",
  "files": [ { "path": "SKILL.md", "purpose": "The skill overview and file manifest (this document)." } ],
  "downloadUrl": "https://www.developersdigest.tech/api/v1/skills/clerk-nextjs/download"
}

-Public and unauthenticated. Unknown slugs return 404 not_found.
-Download the full tree as a zip from /api/v1/skills/{slug}/download; it unpacks cleanly into .claude/skills/.

GET/api/v1/agentsFree

The lean index of the first-party Agent Library: copyable Claude Code subagent definitions.

Response

Field	Type	Req	Description
agents	array	yes	Each { slug, name, description, model, tools }.

curl

curl https://www.developersdigest.tech/api/v1/agents

Response

{
  "agents": [
    { "slug": "code-reviewer", "name": "Code Reviewer", "model": "sonnet", "tools": ["Read", "Grep"] }
  ]
}

-Public and unauthenticated: serves only the curated library, never member-authored agents.
-Fetch the full definition (frontmatter + system prompt) from GET /api/v1/agents/{slug}; save it into .claude/agents/.

POST/api/v1/agents/runMetered

Run one of your managed agents in a sandbox with a single input, buffered to completion. The API sibling of the dashboard run flow.

JSON body

Field	Type	Req	Description
agent	string	yes	Your agent's project id or slug.
input	string	yes	The task to send as the kickoff message.

Response

Field	Type	Req	Description
sessionId	string	yes	Run session id; feed it to GET /api/v1/agents/runs/{id}.
status	string	yes	Final run status.
output	string	yes	The agent's final output.
creditsSpent	number	yes	Credits charged for the sandbox boot (0 when a warm sandbox is reused or for owners).
balance	number	yes	Remaining credit balance.

curl

curl -X POST https://www.developersdigest.tech/api/v1/agents/run \
  -H "Authorization: Bearer dd_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{"agent":"my-agent","input":"Summarize the latest changelog."}'

Response

{
  "sessionId": "run_abc123",
  "status": "completed",
  "output": "Here is the summary...",
  "creditsSpent": 2,
  "balance": 998
}

-Rate limited to 5 runs / 10 minutes per account.
-Runs your own managed agents (built in the dashboard), not the public Agent Library. Unknown agents return 404.
-Sandbox boot plus a buffered run can take minutes; the route caps at 5 minutes. Fetch the event trace afterwards from GET /api/v1/agents/runs/{sessionId}.

MCP Endpoint

Model Context Protocol

The Model Context Protocol endpoint lives at https://www.developersdigest.tech/api/mcp and is served over streamable HTTP via mcp-handler. It exposes the same credit-metered capabilities as the REST API as MCP tools, so agents can call them directly.

A dd_live_ API key is REQUIRED as an Authorization: Bearer header. There is no session on the MCP transport; every tool call resolves its owner from the key.

Generation tools are credit-metered: generate_image is 5 credits, generate_voice is 1 credit, and generate_video is priced per model. Everything else - search, memory, and the Library tools (skills, agents, design systems) - is free.

Beyond the tools below, the server also exposes files (list_folders, list_assets, list_files, get_file), Canvas authoring (create_canvas, create_html_asset, and friends), and slideshow tools (list_slideshows, export_slide_deck_pptx). Connect and list tools to see the full surface.

Endpoint

HTTPhttps://www.developersdigest.tech/api/mcp

Config (mcp.json)

mcp.json

{
  "mcpServers": {
    "developers-digest": {
      "type": "http",
      "url": "https://www.developersdigest.tech/api/mcp",
      "headers": {
        "Authorization": "Bearer dd_live_your_key_here"
      }
    }
  }
}

Tools

generate_image5 credits

Text-to-image via Fal.ai FLUX schnell. Persists to your Studio gallery and returns a durable Blob URL plus a dashboard link. Charged only on success.

prompt: stringsize?: enum

generate_voice1 credit

Text-to-speech (mp3) via the AI Gateway. Persists to your Studio gallery and returns a durable audio URL. Charged only on success.

text: stringvoice?: enummodel?: enum

generate_videoPer model

Text-to-video via the AI Gateway (optional imageUrl for image-to-video models). Persists to your Studio gallery and returns a durable video URL. Charged only on success at the chosen model's rate.

prompt: stringmodel?: enumimageUrl?: string

count_tokensFree

Count tokens in a string using OpenAI's o200k_base encoding.

text: string

search_contentFree

Search Developers Digest content (blog, guides, tools, videos, courses).

query: stringlimit?: number

get_daily_briefFree

Return the latest Developers Digest Daily Brief.

get_balanceFree

Return the caller's credit balance and owner status.

save_memoryFree

Save a note or link to your memory canvas.

kind?: enumtitle?: stringbody?: stringurl?: string

list_memoriesFree

List your saved memories, newest first.

limit?: number

search_memoriesFree

Search your saved memories (case-insensitive match over title and body).

q: string

list_skillsFree

List the Skill Library: vetted, copyable SKILL.md entries. Returns slug, name, description, category. Filter by category, then call get_skill.

category?: enum

get_skillFree

Fetch a complete SKILL.md ready to save into .claude/skills/<slug>/SKILL.md, plus metadata. Use a slug from list_skills.

slug: string

get_skill_fileFree

Fetch the raw contents of one file inside a skill (e.g. reference/roles.md). The on-demand tier of progressive disclosure; discover paths via get_skill first.

slug: stringpath: string

list_agentsFree

List the Agent Library: copyable Claude Code subagent definitions. Returns slug, name, description, model, tools.

get_agentFree

Fetch a complete subagent definition ready to save into .claude/agents/<filename>, plus a usage snippet. Use a slug from list_agents.

slug: string

list_design_systemsFree

List the DESIGN.md Library: copyable, machine-readable design contracts. Returns slug, name, description, best-for, and gradient policy.

get_design_mdFree

Fetch a complete DESIGN.md contract ready to save into your project root, plus metadata. Use a slug from list_design_systems.

slug: string

dd CLI

Command line

The dd CLI is a small, zero-dependency command line tool that calls the /api/v1 REST surface with your dd_live_ key, prints result URLs, and links back to the dashboard studio.

It ships as a single compiled binary for macOS, Linux, and Windows; download it from /dashboard/cli. Every command accepts --json for machine-readable output, and the CLI exits non-zero on any error so it composes cleanly in scripts.

Setup

Shell

# 1. Download the binary for your platform at https://www.developersdigest.tech/dashboard/cli
chmod +x ~/Downloads/dd && mv ~/Downloads/dd /usr/local/bin/dd

# 2. Authenticate (create a key at https://www.developersdigest.tech/dashboard/keys)
export DD_API_KEY=dd_live_your_key_here
# or persist it: ~/.config/devdigest/config.json -> { "apiKey": "dd_live_..." }

dd --help

generate image

Generate an image from a prompt. Prints URL, model, credits spent, balance, and the dashboard link.

Shell

dd generate image "packet streams over a filesystem tree" --size landscape --out hero.png

--sizesquare, square_hd, portrait, landscape, wide (default square_hd).
--out file.pngDownload the generated image to a local file.

generate voice

Synthesize speech and save the MP3. Prints model, voice, credits spent, balance, and the dashboard link.

Shell

dd generate voice "Welcome to Developers Digest." --voice nova --out intro.mp3

--voicealloy, echo, fable, onyx, nova, shimmer (default alloy).
--modelopenai/tts-1, openai/tts-1-hd (default openai/tts-1).
--out file.mp3Output path (default voice-<timestamp>.mp3).

generate text

One-shot text generation (non-streaming). Prints the completion plus model, credits spent, and balance.

Shell

dd generate text "Summarize the MCP spec in three bullets."

--modelWhitelisted chat model id (default google/gemini-3.1-flash-lite).

skills

Progressive skill disclosure, mirroring the MCP library tools: list the index, get a manifest, pull the zip.

Shell

dd skills list
dd skills get clerk-nextjs
dd skills pull clerk-nextjs --out clerk-nextjs.zip

--out file.zippull only: where to save the zip (default <slug>.zip). Unzips into .claude/skills/.

agents

The first-party Agent Library: list copyable Claude Code subagents, or get a full definition.

Shell

dd agents list
dd agents get code-reviewer --out code-reviewer.md

--out file.mdget only: save the definition as a .md you can drop into .claude/agents/.

agent-run

Run one of your managed agents with a single task and print the buffered output. Pipe the task on stdin or pass --input.

Shell

dd agent-run my-agent --input "Summarize the latest changelog."

--inputThe task to send. Alternatively pipe it: cat task.txt | dd agent-run my-agent.
--out file.jsonAlso save the full run result as JSON.

agent-logs

Status and event trace for a run session id returned by agent-run.

Shell

dd agent-logs run_abc123 --limit 20

--limitOnly show the last N events.

gallery

List your recent assets as a table (kind, credits, created at, url).

Shell

dd gallery --limit 10

--limitHow many recent assets to list.

credits / whoami

credits prints your universal balance and plan; whoami adds your account id and owner flag. Both confirm your key works.

Shell

dd credits
dd whoami

Errors & Rate Limits

Status codes & throttling

Every error uses the same shape - { error, message } - with a matching HTTP status, so any client can rely on a single format. Some errors add extra fields.

Status	Code	Meaning	Extra fields
400	bad_request	Validation failed (bad JSON, missing or out-of-range field).	-
401	unauthorized	Missing or invalid API key or session.	-
402	insufficient_credits	Not enough credits for this action. Add credits at /pricing.	balance
404	not_found	No resource for that slug or id (skills, agents, run sessions).	-
429	rate_limited	Too many requests. Slow down and retry.	retryAfterSeconds + Retry-After header
500	internal	Unexpected server error. Metered actions do not charge on failure.	-
503	not_configured	The feature is not enabled on the server (missing provider config).	-

Error shape

Response

{
  "error": "insufficient_credits",
  "message": "Not enough credits for this action. Add credits at /pricing.",
  "balance": 1
}

Rate limits

Endpoint	Limit
POST /api/v1/images	10 / 5 min
POST /api/v1/voice	10 / 5 min
POST /api/v1/text	20 / 5 min
POST /api/v1/video	5 / 5 min
POST /api/v1/agents/run	5 / 10 min
GET /api/v1/favicons	30 / 5 min
GET /api/v1/pricing	60 / 5 min
POST /api/v1/uploads	20 / 5 min
MCP generate_image / generate_voice	10 / 5 min

A 429 includes a retryAfterSeconds field and a Retry-After header. Other /api/v1 reads (me, memory, gallery) are not separately throttled.