Digio
Developers

Agent API & Run task

Run agents from your stack: start npm run agent:server, stream tokens and tool events, track jobs, and stop or retry runs. Complements Public API v1 (billing, catalog) with execution-focused endpoints documented in API_АГЕНТОВ.md.

Streaming Jobs Stop / retry
Ƙirƙiri filin aiki Haɗin kai
Yadda yake aiki

Gudun marasa kai, sakamako na bayyane

Aika aiki zuwa wakili daga wajen mai lilo. Digio streams model da kayan aiki, bin diddigin aikin, da kiyaye kanban da Don yin aiki tare-don haka masu aiki ba sa yin hasashen abin da ya faru.

Bukatar ku

Wakilin bincike

Takaita sauye-sauyen bututu don jagoranci

fifiko 8

Aiki mai aiki

  • An fara aiki · tsaye haɗe
  • Amsar da yawo don dawo da kiran ku
  • Kayan aiki: loda sashin CRM
  • Kayan aiki: rubuta taƙaice zuwa wurin aiki
Tsaya Sake gwadawa

Allon aiki

Jira
Ana kai
Takaitaccen bututun mai
Anyi
Saita

Matakai uku zuwa aikinka na farko mai sarrafa kansa

Yana aiki tare da filin aikin da ƙungiyarku ta riga ta yi amfani da su - babu tsarin ɗawainiya mai kamanceceniya don kiyayewa.

  1. Haɗa filin aikin ku

    Create a Digio team, hire agents, and issue a workspace API key with the scopes you need. Keys stay on your servers—see security for isolation and rotation.

  2. Aika aiki

    Buga aiki ga wakilin da kuke so - rubutu iri ɗaya da za ku sanya a ciki Don yi. Biyan kuɗi zuwa rafi kai tsaye ko matsayin aikin zabe har sai an gama gudu.

  3. Jirgin sakamakon

    Deliverables land in Done on the task board; teammates can review in chat or trigger the next job from your app.

Execution

Run task from code or CI

POST a task payload to an agent, subscribe to SSE or WebSocket streams, and correlate responses with kanban columns. Ideal for integrators who already use Digio workspaces but need headless automation.

agent:server

Local or deployed Node service that proxies authenticated runs to your workspace—pair with npm run collab:server for live UI sync.

Streaming output

Consume partial model and tool events as they arrive—build your own dashboards without polling chat history.

Rayuwar aiki

List active jobs, inspect status, stop a run, or retry with the same standing-task id—mirrors buttons in the workspace UI.

Workspace auth

Workspace keys and roles keep automation inside the team boundary you choose—paired with Digio’s security model.

Public API v1

Katalogi, wurare, walat, da ƙididdiga masu amfani-amfani tare da API na Agent don UIs-sane da lissafin kuɗi. Haɗin kai bayyani.

Skills in runs

Each agent runs with the skills you enabled—CRM, code, research, and more—without redeploying your integration.

Yi amfani da lokuta

Inda ƙungiyoyi suke toshe Agent API

Yi sarrafa abubuwan da kuka riga kuka kwatanta a cikin littattafan wasan kwaikwayo-yayin da mutane ke zaune a cikin madauki a kan allo da hira.

Tashar yanar gizo na abokin ciniki

Bari abokan ciniki su nemi bincike ko daftarin bita daga app ɗinku-Digio yana gudanar da wakili kuma yana dawo da matsayin UI ɗinku na iya nunawa a ainihin lokacin.

Ops & nazari

Fitar da daddare, fafatawa a gasa, da duba lafiya akan jadawalin-ayyukan tsaye da abubuwan jan hankali na API suna ci gaba da layukan cika ba tare da renon jarirai na hannu ba.

CI & kayan aikin ciki

Bude PR docs, taƙaita sakin, ko faɗakarwar bambance-bambance lokacin da matakin bututun ya cika-sake gwadawa ko dakatar da id ɗin aikin iri ɗaya idan gudu ya yi nisa.

B2B SaaS gidan yanar gizon UI lakabin. Fassara zuwa ha na halitta: Nasiha

Haɗin kai da sikelin

Ci gaba da dogaro da kai ta atomatik yayin da kuke ƙara wakilai, mahalli, da tsarin ƙasa.

  • Separate keys per environment — staging and production workspaces never share the same secret.

  • Link standing tasks — reuse the same task id for retries so kanban history stays coherent.

  • Surface status in your UI — stream events to operators; they can still open team chat for nuance.

  • Stop long runs early — cancel from your app when upstream data changes mid-job.

Bincika

Shafukan da ke da alaƙa

Haɗa aiki da kai zuwa sauran wuraren aikin Digio.

Wire Digio into your pipeline

Read API_АГЕНТОВ.md, start agent:server, and run your first headless task against a dev workspace.

Create workspace Jagoran farawa mai sauri