Link software
Free connector; inference extra
Usage billing by selected model/provider
- Allowance and restrictions
- Connector itself has no vendor subscription fee; model inference remains separately billed.
Together Link
● Identity checkedAI coding model routing · Official-source review 2026-10-09 UTC
Together Link connects supported coding agents and desktop harnesses to models hosted by Together AI. It offers model selection or per-session routing and reports token counts and costs after proxied sessions.
- Category
- AI coding model routing
- Free access
- The vendor describes the Link connector as free and open source. Hosted inference is separately billed at model-specific token rates; an ongoing free inference allowance is not confirmed.
- API access
- Together API gateway for supported coding harnesses; optional separate provider keys for router/search features.
Is Together Link right for you?
A good fit for
Fits developers who want to compare hosted model costs in an existing coding workflow and can review provider access and resulting code.
Before you choose
Beta requires macOS or Linux, a Together API key and a supported harness. The official page lists Claude Code/Desktop, Codex, ChatGPT Desktop, OpenCode version 2+ and Pi version 0.80.8+. Model inference is usage-based, not a monthly Link subscription; input, output, cache and promotional rates vary by model. Pricing page explicitly uses per-million-token units; Kimi promotional rates expire October 11, 2026. Existing harness subscriptions may remain separate. Desktop modes use separate profiles and CLI modes session settings, according to the vendor. Routing may require an additional Anthropic key; web search may require Exa. Savings, frontier-quality, configuration isolation and rollback claims were not independently tested. No installer was run, profile modified or key supplied.
Workflow and handoff
Review current supported models and per-token rates, restrict the provider key, compare a small authorized coding task, inspect generated changes and session receipts, then verify the documented switch-back procedure.
Pricing and practical limits
Beta requires macOS or Linux, a Together API key and a supported harness. The official page lists Claude Code/Desktop, Codex, ChatGPT Desktop, OpenCode version 2+ and Pi version 0.80.8+. Model inference is usage-based, not a monthly Link subscription; input, output, cache and promotional rates vary by model. Pricing page explicitly uses per-million-token units; Kimi promotional rates expire October 11, 2026. Existing harness subscriptions may remain separate. Desktop modes use separate profiles and CLI modes session settings, according to the vendor. Routing may require an additional Anthropic key; web search may require Exa. Savings, frontier-quality, configuration isolation and rollback claims were not independently tested. No installer was run, profile modified or key supplied.
Review method
Relevant official-source evidence reviewed individually on 2026-10-09 UTC (October 10 in India). Selected product artwork inspected visually. No hands-on vendor application test was performed. Performance, security and plan enforcement were not independently tested.
What it can do
Features & capabilities
Unknown is different from unavailable. Each fact carries its own evidence.
| Capability | Value | Evidence | Checked |
|---|---|---|---|
| Documented workflow | Together Link connects supported coding agents and desktop harnesses to models hosted by Together AI. It offers model selection or per-session routing and reports token counts and costs after proxied sessions. | Facts sourced | 2026-10-09 |
Understand the total cost
Together Link pricing & plans
Hosted inference
Model-specific token rates
Pay as you go; separate harness or search costs may apply
- Allowance and restrictions
- Input/output/cache token pricing varies by model; check current official table and promotion expiry.
Alternatives to Together Link
View all ↗
RunInfraRunInfra provides hosted open models through OpenAI- and Anthropic-compatible APIs. It offers prepaid token billing and workspace coding plans, with session-aware caching and separately reported cached-token usage.
Explore tool
NumaClawNumaClaw provisions hosted instances for AI agents and company-style agent teams, with templates, model credentials and communication-channel setup.
Explore tool
chat-recallchat-recall stores coding conversations for full-text and knowledge-graph retrieval. It offers MCP access, analytics and code-related findings, with hosted and self-hosted deployment options.
Explore toolThe practical questions
Frequently asked questions
Was this product hands-on tested?
No. The listing uses dated official-source evidence and visual artwork review. Unconfirmed pricing, allowances and integrations are stated explicitly.
What is Together Link used for?
Together Link connects supported coding agents and desktop harnesses to models hosted by Together AI. It offers model selection or per-session routing and reports token counts and costs after proxied sessions.
Does Together Link have a free plan?
The vendor describes the Link connector as free and open source. Hosted inference is separately billed at model-specific token rates; an ongoing free inference allowance is not confirmed.. This record does not confirm an ongoing free plan.
How much does Together Link cost?
Free connector; inference extra. See the billing terms below. See the plan cards for entitlements, billing commitments and seat minimums.
Can I use Together Link through an API?
Together API gateway for supported coding harnesses; optional separate provider keys for router/search features.. API access and subscription access may have different terms; consult the linked sources.
What should I check before choosing it?
Beta requires macOS or Linux, a Together API key and a supported harness. The official page lists Claude Code/Desktop, Codex, ChatGPT Desktop, OpenCode version 2+ and Pi version 0.80.8+. Model inference is usage-based, not a monthly Link subscription; input, output, cache and promotional rates vary by model. Pricing page explicitly uses per-million-token units; Kimi promotional rates expire October 11, 2026. Existing harness subscriptions may remain separate. Desktop modes use separate profiles and CLI modes session settings, according to the vendor. Routing may require an additional Anthropic key; web search may require Exa. Savings, frontier-quality, configuration isolation and rollback claims were not independently tested. No installer was run, profile modified or key supplied.
Price history
No retained pricing changes yet. A current price alone does not establish a historical trend.
How this profile is supported
Official product information
Product identity and relevant official-source evidence reviewed individually.
Read original source ↗Costs and constraints
Beta requires macOS or Linux, a Together API key and a supported harness. The official page lists Claude Code/Desktop, Codex, ChatGPT Desktop, OpenCode version 2+ and Pi version 0.80.8+. Model inference is usage-based, not a monthly Link subscription; input, output, cache and promotional rates vary by model. Pricing page explicitly uses per-million-token units; Kimi promotional rates expire October 11, 2026. Existing harness subscriptions may remain separate. Desktop modes use separate profiles and CLI modes session settings, according to the vendor. Routing may require an additional Anthropic key; web search may require Exa. Savings, frontier-quality, configuration isolation and rollback claims were not independently tested. No installer was run, profile modified or key supplied.
Read original source ↗Facts apply to the named version and check date. Send a sourced correction if something changed.